# Tanvo — full text Tanvo (https://tanvo.ai) is an AI video and image studio. Type an idea or add a photo and it makes an image, or a short video with sound, on one of 28 models. Everything runs on one credit balance, the price of a run is shown before it starts, failed runs are refunded, and the first image needs no account. ## Models - Tanvo Video v1 (video, In-house): The house video engine. Fast 480p drafts from a prompt or a photo. Resolutions 480p; lengths 5, 8 s. Default run about 20 credits. https://tanvo.ai/video/studio-video-v1 - MiniMax H3 (video, MiniMax): Quick clips from a prompt or photo, up to 10 seconds at 480p to 1080p. Resolutions 480p, 768p, 1080p; lengths 4, 6, 8, 10 s. Default run about 90 credits. https://tanvo.ai/video/minimax-h3 - Seedance 2.5 (video, ByteDance): Long takes up to 30 seconds with generated audio. Resolutions 480p, 720p, 1080p; lengths 5, 10, 15, 20, 30 s. Default run about 300 credits. https://tanvo.ai/video/seedance-2-5 - Seedance 2.0 (video, ByteDance): Video from text or images with native audio. 480p to 1080p. Resolutions 480p, 720p, 1080p; lengths 4, 6, 8, 10, 15 s. Default run about 480 credits. https://tanvo.ai/video/seedance-2 - Seedance 2 Fast (video, ByteDance): The quicker, cheaper Seedance tier for draft passes. Resolutions 480p, 720p; lengths 4, 6, 8, 10 s. Default run about 445 credits. https://tanvo.ai/video/seedance-2-fast - Seedance 2 Mini (video, ByteDance): The cheapest Seedance tier: 480p or 720p, 4 to 15 seconds, sound included. Resolutions 480p, 720p; lengths 4, 6, 8, 10, 12, 15 s. Default run about 130 credits. https://tanvo.ai/video/seedance-2-mini - Seedance 1.5 Pro (video, ByteDance): Mid-weight text or image to video with optional audio. Resolutions 480p, 720p, 1080p; lengths 4, 6, 8, 10, 12 s. Default run about 115 credits. https://tanvo.ai/video/seedance-1-5-pro - Kling 3.0 (video, Kuaishou): General-purpose video from text or image, 720p to 4K, optional audio. Resolutions 720p, 1080p, 4K; lengths 5, 10 s. Default run about 155 credits. https://tanvo.ai/video/kling-3-0 - Kling 2.6 (video, Kuaishou): Smooth, cinematic 5 or 10 second clips from text or a photo, with optional sound and speech made in the same pass. Resolutions 1080p; lengths 5, 10 s. Default run about 130 credits. https://tanvo.ai/video/kling-2-6 - Kling 3.0 Turbo (video, Kuaishou): The quick version of Kling 3.0: 5 to 15 second clips from text or a start image at 720p or 1080p. Resolutions 720p, 1080p; lengths 5, 10, 15 s. Default run about 210 credits. https://tanvo.ai/video/kling-3-turbo - Veo 3.1 (video, Google): Cinematic video with sound. Lite, Fast and Quality tiers. Resolutions 720p, 1080p; lengths 4, 6, 8 s. Default run about 115 credits. https://tanvo.ai/video/veo-3-1 - Wan 3.0 (video, Alibaba): Up to 30 seconds in one take with audio written in the same pass. Resolutions 480p, 720p, 1080p; lengths 5, 10, 15, 30 s. Default run about 185 credits. https://tanvo.ai/video/wan-3-0 - LTX 2.5 (video, Lightricks): Video with synchronised sound from text or a start image, any length from 5 to 20 seconds, up to 4K. Resolutions 720p, 1080p, 2K, 4K; lengths 5, 10, 15, 20 s. Default run about 185 credits. https://tanvo.ai/video/ltx-2-5 - FLUX 3 (video, Black Forest Labs): Text or image to video with sound, up to 20 seconds at 720p or 1080p. Resolutions 720p, 1080p; lengths 5, 10, 20 s. Default run about 315 credits. https://tanvo.ai/video/flux-3 - Tanvo Image v1 (image, In-house): The house image engine. Clean results from text in seconds. Resolutions 1K. Default run about 5 credits. https://tanvo.ai/image/studio-image-v1 - Nano Banana 2 (image, Google): Generate and edit high-quality images from text or up to 14 photos. Resolutions 1K, 2K, 4K. Default run about 25 credits. https://tanvo.ai/image/nano-banana-2 - Nano Banana 2 Lite (image, Google): The light tier. Generate and edit at one flat rate. Resolutions 1K. Default run about 15 credits. https://tanvo.ai/image/nano-banana-2-lite - Nano Banana Pro (image, Google): The top tier for text to image, with consistent run times. Resolutions 1K, 2K, 4K. Default run about 50 credits. https://tanvo.ai/image/nano-banana-pro - Nano Banana (image, Google): The original Nano Banana (Gemini 2.5 Flash Image): follows plain instructions and edits photos while keeping faces the same. Resolutions 1K. Default run about 15 credits. https://tanvo.ai/image/nano-banana - GPT Image 2 (image, OpenAI): Text to image and description-based edits, 1K to 4K. Resolutions 1K, 2K, 4K. Default run about 25 credits. https://tanvo.ai/image/gpt-image-2 - GPT Image 2.5 (image, OpenAI): The newest OpenAI image tier on the fast track. Resolutions 1K, 2K, 4K. Default run about 10 credits. https://tanvo.ai/image/gpt-image-2-5 - Seedream 5.0 Lite (image, ByteDance): Fast, low-cost stills and edits at a flat price. Resolutions 2K. Default run about 15 credits. https://tanvo.ai/image/seedream-5-lite - Seedream 4.0 (image, ByteDance): Sharp, high-resolution stills up to 4K and multi-photo edits that keep the subject consistent. Resolutions 2K, 4K. Default run about 10 credits. https://tanvo.ai/image/seedream-4 - Seedream 4.5 (image, ByteDance): Seedream 4.0 with clearer lettering and closer prompt following, up to 4K: posters, product shots, brand visuals. Resolutions 2K, 4K. Default run about 15 credits. https://tanvo.ai/image/seedream-4-5 - Seedream 5.0 Pro (image, ByteDance): ByteDance's flagship stills model: dense layouts and infographics, realistic portraits, text in many languages, precise edits. Resolutions 1K, 2K. Default run about 15 credits. https://tanvo.ai/image/seedream-5-pro - Qwen Image 3 (image, Alibaba): Built for readable text and page layout: posters, infographics, storyboards. Resolutions 1K, 2K. Default run about 10 credits. https://tanvo.ai/image/qwen-image-3 - Qwen Image 2.1 (image, Alibaba): Quick, low-cost stills and multi-photo edits, good at lettering including Chinese. Resolutions 1K, 2K. Default run about 10 credits. https://tanvo.ai/image/qwen-image-2-1 - Grok Imagine Image 2.0 (image, xAI): xAI's newer image model: better prompt understanding and lettering than the first Grok Imagine, and photo edits. Resolutions 1K. Default run about 20 credits. https://tanvo.ai/image/grok-imagine-2 ## Apps ### Hotel Lobby AI Video https://tanvo.ai/video/hotel-lobby-ai-video · video · Duets & trends You and anyone else, trading verses at one hanging mic in an orange booth, with sound, in about three minutes. Looks: Classic booth, Wedding day, Rap battle, Matching suits, Halloween, Retro 70s. Q: Which app makes the hotel lobby AI video? A: You do not need an app. Upload two photos here; the generator runs Seedance 2.5 with a prompt that already describes the booth, the mic and the performance, and returns the clip with sound. Q: Why does one of the faces not look like the photo? A: Use a sharp, front-facing photo with the face filling a good part of the frame, and only one person per photo. Side profiles, sunglasses and group shots are the usual culprits. Q: Can I put pets, cartoons or drawings in the booth? A: Yes. Anything with a face works as a reference; the prompt keeps the booth and the staging, and the model adapts the performance to whoever you cast. Q: How long can the clip be? A: 10 to 30 seconds. 10 seconds fits a single swap of lines; 30 seconds gives the full back-and-forth. Q: Does it use the original footage or audio? A: No. Every clip is generated from your photos and the prompt; the sound is generated with it. Add your own track afterwards if you want a specific song. Q: What does it cost? A: The price in credits shows next to the Generate button before you run it and changes with length and resolution. Failed runs are refunded. ### AI Crying Filter https://tanvo.ai/video/ai-crying-filter · video · Reactions & effects One smile, one tear, one slow walk away — the sad-girl-walk clip, from a single photo. Looks: Smile to tears, Holding it in, Grin to full cry, Rain on the window, Laughing through tears. Q: Which app does the AI crying filter? A: No app needed. Upload a photo here; the generator runs Seedance 2.0 with a prompt that already describes the smile, the tear and the walk away. Q: Why is the tear missing or the face changing? A: Use a sharp, front-facing photo with no sunglasses and only one person in frame. Side profiles and group shots are where the face drifts. Q: Can I change where they walk? A: Yes, edit the last sentence of the prompt: a rainy street, a school corridor, a beach at dusk. Q: Can I use the result commercially? A: Runs on a paid plan or credit pack come with a commercial licence. Free runs are for personal use; see the Terms. Q: What does it cost? A: The credit price shows next to the Generate button before you run it. Failed runs are refunded. ### AI Arrest Filter https://tanvo.ai/video/ai-arrest-filter · video · Reactions & effects Walked out in handcuffs, completely unbothered — from one photo. Looks: Walked out at night, Coffee in hand, Escorted from the wedding, Grocery run, Karaoke night. Q: What is the AI arrest filter? A: A one-photo video effect: the person is walked out in handcuffs under flashing lights, unbothered. The joke is the caption you add. Q: Which model makes it? A: Seedance 2.0 from a single reference photo, with generated sound. Q: Can I do two people? A: Yes: switch the model to Seedance 2.5, upload two photos and say both are being escorted. Q: Why do the officers look odd? A: Shorter clips (6s) keep them cleaner; ask for them to stay slightly out of focus behind the subject. Q: How much does it cost? A: Shown next to Generate before you run. Failed runs are refunded. ### Ceiling Fan Transition https://tanvo.ai/video/ceiling-fan-transition · video · Fashion & body Lie under the fan, and every blade sweep swaps your outfit. Looks: Bedroom daylight, Neon night, Sunlit floor, Hotel room. Q: Do I need a ceiling fan? A: No. The fan, the room and the overhead camera are all generated. You only need photos. Q: How many outfits can I show? A: Up to four in one clip: your main photo plus three. Order of upload is order of appearance. Q: Which model is used? A: Seedance 2.5, which takes several reference photos at once and keeps the face and pose consistent between swaps. Q: The transitions are not on the blade. Why? A: Keep the clip at 10 to 15 seconds and three to four outfits; more swaps than blade passes forces the model to cut between them. Q: What does it cost? A: Shown before you run; depends on length and resolution. Failed runs are refunded. ### Genjutsu Video https://tanvo.ai/video/genjutsu-video · video · Duets & trends Keep the moves, the camera and the cuts of a real clip — recast who and where. Looks: Neon car park, Rooftop sunset. Q: What is a genjutsu video? A: A clip-to-clip edit: the model reads your source video, holds its choreography, timing and camera path, and writes a new clip with a different person, outfit or place. Q: Which model powers it? A: Seedance 2.5 with a reference video; it reads the whole clip rather than a single frame. Q: Why is it slow? A: Length drives time: a 5-second edit takes a few minutes, 30 seconds considerably longer. You can leave the page; the result is saved to your history. Q: What clip works best? A: One subject, steady framing, no overlays or captions, under 30 seconds. Q: What does it cost and what if it fails? A: Priced per second of the source clip; the exact number shows before you run. Failed runs are refunded. ### STORM II AI Video https://tanvo.ai/video/storm-ii-ai-video · video · Duets & trends Stand still at the centre while a whole formation moves around you. Looks: Classic, White among suits, Office all-hands, Knights' courtyard. Q: What is STORM II? A: A music video by GENER8ION and Yung Lean, directed by Romain Gavras with choreography by Damien Jalet. The formation shot — one still figure, a moving crowd — is what people recreate. Q: Was the original made with AI? A: No, it was live-action with real dancers. This page generates a new clip from your photo. Q: Why does the crowd look like a blur? A: Ask for fewer ranks and 720p or above; 480p struggles with dozens of small figures. Q: Why does the lead move? A: Keep 'stands perfectly still' in the prompt and avoid describing any action for the subject. Q: What does it cost? A: Priced per second; the number shows before you run. Failed runs are refunded. ### Pata Chalega AI Video https://tanvo.ai/video/pata-chalega-ai-video · video · Duets & trends You, driving a black luxury sedan, smiling and waving through the windshield. Looks: Sunny drive, Night drive, Aviators on the highway, Rain and wipers, Village road, Peace sign. Q: Do I need a photo of a car? A: No. Upload one photo of the person; the car, the street and the camera move are generated. Q: Which model is used? A: MiniMax H3 — fast and cheap for this short, simple shot. Switch to Seedance 2.0 for more detail. Q: Does it include the song? A: No. Add the Pata Chalega audio in Reels or Shorts when you post; the clip comes with ambient sound only. Q: Can I make it longer? A: Up to 10 seconds on this model, but the trend's hook fits 6. Q: What does it cost? A: Shown before you run; this is one of the cheapest clips on the site. Failed runs are refunded. ### AI Laughing Face Effect https://tanvo.ai/video/ai-laughing-face-effect · video · Reactions & effects A still photo laughs on cue — the reaction clip, from one photo. Looks: Classic, Trying not to laugh, Reading the group chat, Baby giggle. Q: What is this effect? A: A one-photo video where the person reacts — laughing — on cue, with the face kept from the photo. Q: Which model? A: Seedance 2.0 from one reference image, with generated sound. Q: Can I chain reactions? A: Yes: make the laughing, shocked and scared clips from the same photo and cut them together. Q: Why is the expression weak? A: Start from a neutral, well-lit photo and keep the camera locked; the model has more room to perform. Q: What does it cost? A: Shown before you run. Failed runs are refunded. ### AI Shocked Face Effect https://tanvo.ai/video/ai-shocked-face-effect · video · Reactions & effects A still photo gasps on cue — the reaction clip, from one photo. Looks: Classic, Dramatic zoom, Double take, Mind blown. Q: What is this effect? A: A one-photo video where the person reacts — shocked — on cue, with the face kept from the photo. Q: Which model? A: Seedance 2.0 from one reference image, with generated sound. Q: Can I chain reactions? A: Yes: make the laughing, shocked and scared clips from the same photo and cut them together. Q: Why is the expression weak? A: Start from a neutral, well-lit photo and keep the camera locked; the model has more room to perform. Q: What does it cost? A: Shown before you run. Failed runs are refunded. ### AI Scared Face Effect https://tanvo.ai/video/ai-scared-face-effect · video · Reactions & effects A still photo panics on cue — the reaction clip, from one photo. Looks: Flickering light, Something off camera. Q: What is this effect? A: A one-photo video where the person reacts — scared — on cue, with the face kept from the photo. Q: Which model? A: Seedance 2.0 from one reference image, with generated sound. Q: Can I chain reactions? A: Yes: make the laughing, shocked and scared clips from the same photo and cut them together. Q: Why is the expression weak? A: Start from a neutral, well-lit photo and keep the camera locked; the model has more room to perform. Q: What does it cost? A: Shown before you run. Failed runs are refunded. ### Split Face Parents https://tanvo.ai/image/split-face-parents · image · Photoshoots & family Half your face, half your parent's — lined up into one portrait. Looks: Studio split, Thank-you B&W, Then and now, Sibling split, Golden window light. Q: What is the split face trend? A: One portrait made of two faces: the left half from one photo, the right half from another, aligned along the centre line. Q: Which photos work best? A: Both front-facing, eyes level, similar lighting. Old scanned photos work if the face is sharp. Q: Can I do more than two people? A: Yes with a three-way split, but it reads best with two. Edit the prompt to 'thirds' if you want to try. Q: Which model is used? A: Nano Banana 2, which takes several reference photos and is good at keeping both faces faithful. Q: What does it cost? A: Shown before you run. Failed runs are refunded. ### AI Halloween Costume Generator https://tanvo.ai/image/ai-halloween-costume-generator · image · Face filters Try a witch, vampire or werewolf on your own face before buying anything. Looks: Witch, Y2K vampire, Skeleton face paint, Scarecrow, 1920s flapper, Retro astronaut, Werewolf, Forest fairy. Q: How does it work? A: One photo plus a costume description. The model keeps your face and body and dresses you, with props and a scene. Q: Can I do group or couple costumes? A: Yes: switch to two photos and describe both costumes; say who wears what. Q: Will it look like me? A: Yes if the photo is front-facing and sharp. Sunglasses, hats and group shots reduce likeness. Q: Can I use the result as a reference for a real costume? A: That is the idea. Download it and take it shopping or to the tailor. Q: What does it cost? A: Around 25 credits per 1K image. Failed runs are refunded. ### AI LARP Generator https://tanvo.ai/image/ai-larp-generator · image · Characters & fandom One selfie; a supercar, a private jet and a penthouse you do not own. Looks: Supercar & jet, Penthouse sunset, Yacht deck, First-class suite, Rooftop dinner, Ski chalet. Q: What is a LARP photo? A: A staged 'rich life' photo — car, jet, penthouse — generated from an ordinary selfie, posted as a joke or as an obvious flex. Q: Will my face look right? A: Yes with a sharp, front-facing selfie. The model keeps the face and body and builds the scene around them. Q: Can I add a friend? A: Yes: two photos, and say both are in the shot. Q: Can I use it commercially? A: Runs on a paid plan or credit pack come with a commercial licence. Free runs are for personal use; see the Terms. Q: What does it cost? A: About 25 credits per 1K image. Failed runs are refunded. ### What Animal Do I Look Like? https://tanvo.ai/image/what-animal-do-i-look-like · image · Memes & cards Upload a selfie and meet the animal that shares your face. Looks: Spirit animal card, Wildlife magazine, Bird edition, Friend group, Royal oil portrait. Q: How does it pick the animal? A: From the face: shape, eyes, brows, expression and hair. It is a resemblance call, not a personality quiz. Q: Can I choose the animal? A: Yes, replace 'decide which animal' in the prompt with the animal you want. Q: Does it keep my photo? A: Your original appears on the left of the result unchanged; the right side is generated. Q: Which model is used? A: Nano Banana 2, which handles the side-by-side layout and the caption in one pass. Q: What does it cost? A: About 25 credits per image. Failed runs are refunded. ### AI Bob Haircut Filter https://tanvo.ai/image/ai-bob-haircut-filter · image · Face filters Chin bob, French bob or a lob — on your own face, before the chop. Looks: Sleek chin bob, French bob, fringe, Italian bob, Wavy lob, Bixie crop, Platinum blunt bob. Q: Does it change anything but the hair? A: No. The prompt locks the face, skin, clothes, background and light; only the hairstyle is edited. Q: Can I change the colour as well? A: Yes, add it: 'copper', 'platinum blonde', 'jet black'. Q: Which model does the editing? A: Nano Banana 2, which is built for edits that touch one thing and leave the rest. Q: Why does the hair look pasted on? A: Use a photo with even light and the hair fully visible; very dark or backlit photos give flat results. Q: What does it cost? A: About 25 credits per 1K image. Failed runs are refunded. ### Italian Brainrot Generator https://tanvo.ai/image/italian-brainrot-generator · image · Memes & cards Invent your own absurd creature in the Tralalero Tralala style — and name it. Looks: Flamingo moka pot, Seaplane pelican, Granita hippo, Scooter chameleon, Pizza tortoise, Gondola swan. Q: What is Italian brainrot? A: A meme genre of absurd photorealistic hybrid creatures with Italian-sounding rhyming names, usually narrated in a flat TTS voice. Q: Why GPT Image 2.5? A: It renders photoreal hybrids and legible caption text in one image at a low price. Switch models if you prefer a different look. Q: Can I animate it? A: Yes: open the result in the video generator with Seedance 2.0 and describe the creature walking through the piazza. Q: Can I use it commercially? A: Results on a paid plan come with a commercial licence; avoid using existing characters' names. Q: What does it cost? A: About 10 credits per 1K image. Failed runs are refunded. ### AI Figurine Generator https://tanvo.ai/image/ai-figurine-generator · image · Characters & fandom Turn yourself into a boxed collectible figurine on a desk. Looks: Desk collectible, Chibi vinyl toy, Action figure pack, Pet figurine gift, Tabletop miniature, Anime scale figure. Q: What is the figurine trend? A: A photo-to-image effect that renders you as a collectible PVC figure on a desk, with a retail box and a 3D model on screen. Q: Will it keep my face? A: Yes, slightly stylised like a real figure. Use a sharp, front-facing photo. Q: Which model is used? A: Nano Banana 2, which handles the figure, box art and scene in one pass. Q: Can I get a 3D file? A: No, this is an image. The monitor in the scene is part of the picture. Q: What does it cost? A: About 25 credits per 1K image. Failed runs are refunded. ### 90s Yearbook Photos https://tanvo.ai/image/90s-yearbook-photos · image · Photoshoots & family The laser-background school portrait you never had. Looks: Laser backdrop, Grunge '94, Mall glamour shot, 80s prom, Varsity athlete, Y2K digicam. Q: What makes it look 90s? A: The laser studio backdrop, feathered or gelled hair, denim or chunky knits, the closed-mouth smile and the film grain — all written into the prompt. Q: Will it look like me? A: Yes; the face is kept from the photo, the hair and clothes are changed. Q: Can I do other decades? A: Yes, edit the prompt: 80s glamour shot, 70s school portrait, 2000s MySpace. Q: Which model? A: Nano Banana 2, which keeps faces faithful while restyling everything else. Q: What does it cost? A: About 25 credits per 1K image. Failed runs are refunded. ### AI Outfit Generator https://tanvo.ai/image/ai-outfit-generator · image · Fashion & body Try any outfit on your own photo — describe it or upload it. Looks: Outfit from photo, Old money, Streetwear, Wedding guest, Interview ready, Winter layers. Q: Does it keep my face and pose? A: Yes. The prompt locks face, hair, body shape, pose, background and light; only the clothing is edited. Q: Can I use a real product photo? A: Yes, upload it as the second photo and say 'wear the outfit from the second photo'. Q: Which model? A: Nano Banana 2, which is built for edits that change one thing and leave the rest of the photo untouched. Q: Why does the fit look off? A: Use a photo where the body is fully visible and not cropped at the edges; loose or layered garments need a clear silhouette to work from. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### AI Clothes Swap https://tanvo.ai/image/ai-clothes-swap · image · Fashion & body Swap the clothes between two photos, or put a garment from one onto the other. Looks: Full outfit, Just the jacket, Swap with a friend, Lookbook shot. Q: Which photo goes where? A: First photo is the person who ends up wearing the clothes; second is the source of the clothes. Q: Can I swap only a jacket? A: Yes, say so: 'only the jacket from the second photo; keep the rest of the outfit'. Q: Which model? A: Nano Banana 2, which is built for edits that change one thing and leave the rest of the photo untouched. Q: Does it work with patterns and logos? A: Mostly yes; very small text on garments may blur. Use a 2K output for prints. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### AI Gender Swap https://tanvo.ai/image/ai-gender-swap · image · Face filters See yourself as the other gender, same face, same photo. Looks: Natural swap, Me and my twin, 1950s studio portrait, Street-style candid. Q: Will it still look like me? A: That is the goal: the prompt keeps identity, age, expression and skin tone and changes the gendered features. Q: Can I do several people at once? A: One at a time gives the best likeness; group photos tend to drift. Q: Which model? A: Nano Banana 2, which is built for edits that change one thing and leave the rest of the photo untouched. Q: Is my photo used for anything else? A: No. It is stored with your result so you can see what it was made from, and deleted with it. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### AI Aging Filter https://tanvo.ai/image/ai-aging-filter · image · Face filters See yourself at 60, 70, 80 — or 20 years younger. Looks: At 70, Five decades, 20 years younger, Me and future me, At 50. Q: How realistic is it? A: Very, when the source photo is sharp and front-facing. It changes what time changes — skin, hair, jawline — and leaves identity alone. Q: Can it make me younger? A: Yes: 'make the person look 25' or 'remove 20 years'. Q: Which model? A: Nano Banana 2, which is built for edits that change one thing and leave the rest of the photo untouched. Q: Why does it look like a different person? A: Usually a low-resolution or side-angle photo. Use a clear frontal portrait. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### AI Baby Generator https://tanvo.ai/image/ai-baby-generator · image · Photoshoots & family What would our baby look like? Two photos, one answer. Looks: One-year-old, Newborn swaddle, Toddler in the garden, First day of school, Family portrait. Q: Is this genetically accurate? A: No. It is a visual blend of two faces, made for fun. Q: Can I choose boy or girl, or an age? A: Yes, say it in the prompt. Q: Which model? A: Nano Banana 2, which takes two reference photos and blends features convincingly. Q: Is it safe for an announcement post? A: Yes; the result is a generated image of a child who does not exist. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### AI Tattoo Generator https://tanvo.ai/image/ai-tattoo-generator · image · Fashion & body Design the tattoo first — style, placement, size — then show it to your artist. Looks: Fine-line botanical, American traditional, Geometric blackwork, Script lettering, Watercolour splash, From your sketch. Q: Can a tattoo artist use this? A: Yes as a reference. Most artists will redraw it in their own hand; bring the design and talk about placement and size. Q: Why this model? A: GPT Image 2.5 draws clean linework and readable lettering, which matter more here than photorealism. Q: Can I see it on my body? A: Yes, the tattoo try-on page takes your photo and a design. Q: Is the design mine? A: Results on a paid plan come with a commercial licence; it is yours to get inked. Q: What does it cost? A: About 10 credits per 1K image. Failed runs are refunded. ### AI Muscle Generator https://tanvo.ai/image/ai-muscle-generator · image · Fashion & body See the after photo before the gym: realistic muscle on your own body. Looks: One year of training, Lean and toned, Gym mirror selfie, Before and after, Hero costume fitting, Strong at seventy. Q: Does it keep my face? A: Yes; identity, face, height and pose are locked. Only body composition changes. Q: Can it do the opposite? A: Yes, 'slimmer' or 'leaner' instead of 'more muscular'. Q: Which model? A: Nano Banana 2, which is built for edits that change one thing and leave the rest of the photo untouched. Q: Why do the clothes look wrong? A: Loose clothing hides the change; fitted tops and shorts give the model something to work with. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### AI Selfie Generator https://tanvo.ai/image/ai-selfie-generator · image · Photoshoots & family New selfies of you — different place, outfit, light — from one photo. Looks: Sunny Lisbon street, Mirror outfit check, Rainy night city, Golden-hour beach, Festival day, Café window light. Q: Will it look like a real selfie? A: The prompt asks for phone-camera framing and lens look; keep the scene plausible and it passes. Q: Can I include a friend? A: Yes, two photos and 'both in the selfie'. Q: Which model? A: Nano Banana 2, which keeps the face faithful while building the scene. Q: Should I disclose it is AI? A: Platforms increasingly require it; check the rules where you post. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### AI Full Body Generator https://tanvo.ai/image/ai-full-body-generator · image · Photoshoots & family Turn a headshot into a full-body photo you can use anywhere. Looks: Studio smart-casual, Street style walk, Business full-length, Character turnaround, Summer lookbook. Q: Will the body match the face? A: The model matches skin tone, apparent age and build; say 'tall', 'petite' or similar if you want to steer it. Q: Can I set the pose? A: Yes, describe it, or add a second photo as a pose reference. Q: Which model? A: Nano Banana 2, which is built for edits that change one thing and leave the rest of the photo untouched. Q: Can I use it commercially? A: On a paid plan, yes, subject to the rights in the source photo. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### HD Photo Converter https://tanvo.ai/image/hd-photo-converter · image · Photo fixes Sharpen, de-noise and upscale a blurry or low-res photo to 4K. Looks: Sharpen and upscale, Crisp text and edges, Sharpen and relight. Q: Does it change the photo? A: No new content: the prompt restricts it to sharpening, de-noising and detail recovery. Q: How big can the output be? A: Up to 4K on this model. Q: Which model? A: Nano Banana 2, used in restore mode. For very old damaged prints use the old photo restoration page. Q: Why does a face look slightly different? A: Heavy blur leaves the model guessing. Try 2K before 4K, and start from the largest original you have. Q: What does it cost? A: 25 credits at 1K, 40 at 2K, 60 at 4K. Failed runs are refunded. ### Add Person to Photo https://tanvo.ai/image/add-person-to-photo · image · Photo fixes Put someone into a photo they missed, with matching light and scale. Looks: Beside the group, The photographer, in, Add the family pet. Q: Will it look real? A: With a clear photo of the person and a sensible placement, yes. The prompt matches light, scale and grain. Q: Can I add a deceased loved one? A: Yes, from an old photo. Sharp originals work best; scanned prints are fine. Q: Which model? A: Nano Banana 2, which takes two references and preserves the original photo exactly. Q: Can I add more than one person? A: Yes, up to several reference photos; add them one at a time for the cleanest result. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### Remove Person From Photo https://tanvo.ai/image/remove-person-from-photo · image · Photo fixes Take someone out and fill the background like they were never there. Looks: Remove the photobomber, Leave just one, Clear the clutter. Q: Can it remove several people? A: Yes, list them, or say 'everyone except the two in the centre'. Q: What about objects? A: Same page, same prompt: 'remove the bin on the left'. Q: Which model? A: Nano Banana 2, which is built for edits that change one thing and leave the rest of the photo untouched. Q: Why is the fill blurry? A: Large removals on busy backgrounds are hard; use a 2K output and describe what should be there. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### AI Product Photo Editor https://tanvo.ai/image/ai-product-photo-editor · image · Products & business Clean backgrounds, studio light and lifestyle scenes for any product shot. Looks: White studio, Lifestyle scene, Colour-block ad. Q: Does it change the product? A: No; the prompt keeps shape, labels and details and only changes the surroundings and light. Q: Can I remove the background only? A: Yes: 'transparent background, no shadow' and PNG output. Q: Which model? A: Nano Banana 2, which is built for edits that change one thing and leave the rest of the photo untouched. Q: Is it good enough for a marketplace? A: For most categories yes at 2K; check the platform's background rules. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### Ghibli-Style AI Generator https://tanvo.ai/image/ghibli-ai-generator · image · Styles Your photo as a hand-painted anime film frame — soft light, painted skies. Looks: Painted afternoon, Pet in a meadow, Family on a hill, Rainy lantern lane, Cosy kitchen breakfast, Snowy village. Q: Is this an official Studio Ghibli product? A: No. This page applies a hand-painted anime film look to your photo; it is not affiliated with any studio. Q: Will faces still look like us? A: Yes, stylised: the prompt keeps composition, people and expressions recognisable. Q: Which model? A: Nano Banana 2, which handles style transfer on photos while keeping subjects intact. Q: Can I animate it? A: Yes: open the result in the video generator and describe gentle motion — wind, drifting clouds, a slow pan. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### Pixar-Style Filter https://tanvo.ai/image/pixar-filter · image · Styles Your photo as a 3D animated film character — big eyes, soft lighting, cinematic render. Looks: 3D character, Pet sidekick, Movie poster, Animated family, Little chef, Space explorer. Q: Is this affiliated with Pixar? A: No. This page applies a 3D animated-film style; it is not connected to any studio. Q: Will it look like me? A: Yes, stylised: identity, hairstyle, expression and clothing are kept. Q: Which model? A: Nano Banana 2. Q: Can I do a group? A: Yes, but one person per run then combine, or accept some drift in large groups. Q: What does it cost? A: About 25 credits per 1K image, shown before you run. Failed runs are refunded. ### AI Dance Move Generator https://tanvo.ai/video/ai-twerk-generator · video · Motion & camera A still photo does the viral dance — hips, bounce and all — from one upload. Looks: Classic, Kitchen shuffle, Disco funk, The robot. Q: Is this allowed? A: Dance clips with the person fully clothed as photographed, yes. Sexualised content is blocked by moderation and against the terms. Q: Which model? A: Seedance 2.0 from one reference photo, with generated music. Q: Can I choose the song? A: The clip comes with generated music; add your own track in the editor when you post. Q: Why are the legs odd? A: Use a photo with the whole body and feet visible, standing straight on. Q: What does it cost? A: Shown before you run. Failed runs are refunded. ### Animate Illustrations https://tanvo.ai/video/animate-illustrations · video · Motion & camera Picture-book pages and drawings that move while the lines and words stay put. Looks: Classic, Character moves, Snowfall, Book trailer. Q: Will the style change? A: The prompt locks line quality, colours, composition and text; motion is added on top. Q: Which model? A: Seedance 2.0 from one reference image. Seedance 2.5 if you want longer takes. Q: Can I animate a child's drawing? A: Yes, and it is one of the best uses; crayon and marker textures hold up well. Q: Does text stay readable? A: Printed text in the image is preserved; keep the camera locked for best results. Q: What does it cost? A: Shown before you run. Failed runs are refunded. ### Video to Cartoon Converter https://tanvo.ai/video/video-to-cartoon-converter · video · Styles Turn a real clip into animation — same motion, same cuts, cartoon look. Looks: Classic, Anime, Claymation, Comic book. Q: How does it keep the motion? A: Seedance 2.5 reads the whole clip as a reference and rewrites it in the new style at the same length. Q: Which styles work? A: Flat 2D, anime, claymation, comic halftone, paper cut-out. Name it in the prompt. Q: How long does it take? A: Minutes, scaling with clip length. Start with 5–10 seconds. Q: Can I keep the original audio? A: The result has generated sound; lay the original audio back in your editor if you want it. Q: What does it cost? A: Priced per second of the clip; shown before you run. Failed runs are refunded. ### Photo to Video https://tanvo.ai/video/bring-photos-to-life · video · Motion & camera Make a still photo move, blink and breathe while the camera stays locked. Looks: Blink and smile, A warm embrace. ### AI Lip Sync https://tanvo.ai/video/ai-lip-sync · video · Motion & camera One portrait and a voice track — the face speaks every word. Looks: Talk to camera, Lively storyteller. ### Sketch to Video https://tanvo.ai/video/sketch-to-video-ai · video · Motion & camera A pencil drawing starts to move and stays a pencil drawing. Looks: Classic, Pencil test, Sketchbook on a desk, Off the page. ### Bullet Time Effect https://tanvo.ai/video/bullet-time-effect · video · Motion & camera Freeze the moment and orbit the camera around it. Looks: Classic, Skate kickflip, Court dunk, Dog catch. ### Movie Scene Generator https://tanvo.ai/video/ai-movie-scene-generator · video · Motion & camera One still becomes a three-shot sequence. Looks: Rainy diner, Sci-fi hangar, Frontier town, Candlelit hallway. ### ASMR Video Generator https://tanvo.ai/video/ai-asmr-video-generator · video · Motion & camera Macro close-ups with the sound that goes with them. Looks: Glass apple, Hydraulic press, Soap carving, Sugar crust crack. ### Walking Effect https://tanvo.ai/video/ai-walking-effect · video · Motion & camera A still photo walks toward the camera. Looks: Toward the camera, Runway walk, Slow-mo city walk, Family walk. ### Pet Video Generator https://tanvo.ai/video/ai-pet-video-generator · video · Pets Turn one photo of your dog or cat into a clip with sound. Looks: Curious look, Talking pet, Disco dance, Pet birthday. ### Image to Live Wallpaper https://tanvo.ai/video/image-to-live-wallpaper · video · Motion & camera One still, turned into a seamless ambient loop. Looks: Ambient drift, Cinemagraph portrait, Storybook scene, Gentle snowfall. ### Video Extender https://tanvo.ai/video/ai-video-extender · video · Motion & camera Continue a clip past its last frame. Looks: Keep rolling, Give it an ending. ### Dance Video Generator https://tanvo.ai/video/ai-dance-video-generator · video · Motion & camera A full-body photo learns a routine. Looks: Easy groove, Neon rooftop, Dance practice room, Retro disco. ### Product Spin Shot https://tanvo.ai/video/product-spin · video · Products & business A clean rotating hero shot from one product photo. Looks: Dark turntable, White studio, Floating spin, Sunlit orbit. ### Old Photo Restoration https://tanvo.ai/image/old-photo-restoration · image · Photo fixes Repair scratches, fading and blur in a family photo. Looks: Restore only, Restore and colourise, Like a fresh print. ### Hairstyle Changer https://tanvo.ai/image/ai-hairstyle-changer · image · Face filters Try a bob, curls or a buzz cut on your own face first. Looks: Chin-length bob, Shag with bangs, Buzz cut, Textured crop, Butterfly layers, Copper colour. ### Product Mockup Generator https://tanvo.ai/image/ai-product-mockup-generator · image · Products & business Put your product in a studio scene without the studio. Looks: Stone plinth, Floating hero, Water splash, Editorial flat lay, Sunlit lifestyle. ### Pet Portrait Generator https://tanvo.ai/image/ai-pet-portrait-generator · image · Pets A painted portrait of the one who runs the house. Looks: Renaissance oil, Royal monarch, 3D animated, Watercolour, Pop art, Space explorer. ### Sticker Generator https://tanvo.ai/image/ai-sticker-generator · image · Characters & fandom Die-cut stickers from a line of text. Looks: Coffee cloud, Kawaii sheet, Retro badge, Holographic foil, Chat reaction, Hand-drawn doodle. ### Photo to Coloring Page https://tanvo.ai/image/photo-to-coloring-page · image · Crafts & print Clean black outlines, ready to print. Looks: Clean outlines, Big and simple, Detailed patterns, Cartoon redraw. ### Movie Poster Generator https://tanvo.ai/image/ai-movie-poster-generator · image · Memes & cards A one-sheet with a title you can actually read. Looks: Night thriller, Horror, Rom-com, 80s action, Sci-fi epic, Family animation. ### Pixel Art Converter https://tanvo.ai/image/pixel-art-converter · image · Styles Turn any picture into crisp 16-bit pixels. Looks: 16-bit, 8-bit retro, Four-shade green, Game sprite, Isometric diorama, Neon night. ### Tattoo Try-On https://tanvo.ai/image/tattoo-try-on · image · Fashion & body See the design on your skin before the needle. Looks: Healed forearm, Fresh ink, Colour fill, Upper arm, Side of neck. ### Watercolor Generator https://tanvo.ai/image/ai-watercolor-generator · image · Styles Soft washes and paper grain from a photo or a line. Looks: Harbour morning, Botanical study, Ink and wash, Loose portrait, Misty mountains, Storybook. ### Album Cover Generator https://tanvo.ai/image/ai-album-cover-generator · image · Memes & cards A square that looks like it has a sound. Looks: Empty pool, Y2K chrome, Bedroom pop, Night drive, Surreal, Mid-century jazz. ### Text Removal https://tanvo.ai/image/ai-text-removal · image · Photo fixes Take captions and watermarks of your own images out cleanly. Looks: Remove captions, All text, signs too, Replace the caption. ### AI Halloween Photo Trend https://tanvo.ai/image/ai-halloween-photo-trend · image · Photoshoots & family One ordinary photo becomes six film-grain Halloween scenes: porch light, pumpkin pile, something in the doorway behind you. Still unmistakably you. Looks: Porch at night, Pumpkin pile, Something in the doorway, Candlelit study, Lantern at the gate, Varsity jacket, pumpkin patch. Q: Will it still look like me? A: Yes. Every scene keeps your face, hair and expression; only the setting, the lighting and the film look change. Q: Can I add a costume or a pet? A: Type it in the “Make it yours” box, for example “wearing a witch hat” or “with my black cat”. It is added to the prompt for you. Q: Is it free? A: Each image costs a few credits. New accounts get free credits at sign-up and more every day from check-ins, enough for several scenes. Q: Which model does it use? A: Nano Banana 2 by default, which is best at keeping faces. Nano Banana 2 Lite is cheaper; Nano Banana Pro is sharper. Q: Can I do a group photo? A: Yes, upload a photo with everyone in frame. Each person is kept. Q: What size do I get? A: 1K by default, up to 4K. For a phone post, 1K is plenty. ### AI Pumpkin Stencil Generator https://tanvo.ai/image/ai-pumpkin-stencil-generator · image · Crafts & print Upload a face or a pet, get a clean cut-ready stencil, then preview it carved and lit before you pick up the knife. Looks: Portrait stencil, Dog stencil, Pet with a moon, Carved preview, lit, Carved dog, lit, With a BOO banner. Q: Will the stencil actually be cuttable? A: The prompts ask for bold connected shapes with no floating islands. Check the preview; if a piece is isolated, regenerate with “simpler, fewer islands” in the box. Q: What size does it print at? A: Print at 100% on a normal sheet. A medium pumpkin takes a stencil about 15 cm wide; scale in your print dialog if yours is bigger. Q: Does it work with pets? A: Yes. Dogs and cats with clear outlines come out best. Use a photo where the ears and face are not hidden. Q: Can I add text? A: The banner look adds a short word. Keep it to three or four letters; long words do not survive carving. Q: Is it free? A: A stencil costs a few credits and new accounts start with free credits. Q: What is the carved preview for? A: To see if the likeness survives as a lit silhouette before you spend an hour carving. ### AI Zombie Filter https://tanvo.ai/image/ai-zombie-filter · image · Face filters Six stages of undead on your own face: grey skin, clouded eyes, veins up the neck, and friends who still recognise you in the group chat. Looks: Just bitten, Classic walker, Party makeup, Zombie bride, Long gone, Night selfie. Q: Will people still recognise me? A: Yes. The prompts keep face shape, features and hair; only skin, eyes and clothing change. Pick “just bitten” for the most recognisable version. Q: Is there blood? A: No. Every look is movie makeup without gore, so it stays postable everywhere. Q: Can I keep my glasses? A: Add “keep my glasses” in the “Make it yours” box. Q: Can I turn a group photo? A: Yes. Everyone in the frame turns. Q: How many credits? A: A few per image. Free credits at sign-up cover several stages. Q: Can I make a video of it? A: Yes, the Zombie AI Video app makes a 10-second clip from two photos. ### Zombie AI Video https://tanvo.ai/video/zombie-ai-video · video · Duets & trends The zombie trend clip with you two in it: one of you turned, the other holding on, and a flashback to a better day. Ten seconds, with sound. Looks: Field at dusk, Kitchen table, They open their arms, Rainy car park, Pet version, Last bus. Q: Does it keep both faces? A: Yes. The first photo is the survivor, the second the one who turned, and both faces are kept, including on the zombie. Q: How much does it cost? A: From about 220 credits for ten seconds at 480p on Seedance 2 Mini. The table in the generator shows every model and resolution. Q: Can the second photo be a pet? A: Yes, the pet scene is written for a dog; cats work too. Q: Is it gory? A: No. The prompts ask for makeup-style zombies with no blood, so it posts anywhere. Q: Can I change the ending? A: Pick “They open their arms” or type your own ending in the “Make it yours” box. Q: How long does it take? A: Two to four minutes. You can leave the page; it is saved to your history. ### AI Baby Photoshoot https://tanvo.ai/image/ai-baby-photoshoot · image · Photoshoots & family Six studio-grade shoots from the one clear photo you already have: pumpkin patch, cake smash, newborn basket, Christmas. Same face, no studio, no nap schedule. Looks: Pumpkin patch, First-birthday cake smash, Newborn basket, Christmas, Little pumpkin costume, Spring meadow. Q: Will it look like my baby? A: Yes. The prompts keep the face and expression; only the setting, outfit and light change. Q: Can I add a name or age sign? A: Yes, type it in the “Make it yours” box, for example “a wooden sign that says ONE”. Q: Is my baby's photo stored? A: It is stored with the result so you can see what it was made from, and deleted with it. It is never used for anything else. Q: What photo works best? A: Front-on, well lit, no hands or hats covering the face. Q: Can I print it? A: Generate at 2K or 4K for prints up to A4. Q: How much is it? A: A few credits per image. Free credits at sign-up cover a few shoots. ### AI Family Photo Generator https://tanvo.ai/image/ai-family-photo-generator · image · Photoshoots & family Upload the snapshot you already have of everyone, pick a scene, and get the family photo for the holiday card. Every face kept. Looks: Fall pumpkin patch, Christmas pyjamas, Beach at sunset, Studio portrait, Denim at golden hour, 1980s portrait studio. Q: Does it keep everyone's face? A: Yes, and the number of people. The prompt tells the model to keep every face and the same heights. Q: Can I add someone who is not in the photo? A: Use the Add Person to Photo app first, then bring that result here. Q: Can I include our dog? A: If the dog is in the photo it stays. If not, type “add our beagle” in the box. Q: What photo works best? A: Everyone facing the camera, daylight, no sunglasses. Q: Can I print the result? A: Generate at 2K or 4K for cards and prints. Q: How much is it? A: A few credits per image; new accounts start with free credits. ### Hoco Sign Generator https://tanvo.ai/image/hoco-sign-generator · image · Memes & cards Sour candy, a volleyball, a pizza box that says it is cheesy: preview the whole hoco sign, with their name on it, before a single evening of glitter glue. Looks: U + ME = HOCO?, Cheesy pizza box, Volleyball set, Kick it, Glow sticks, Sour candy. Q: Can it put their actual name on the sign? A: Yes. Type the name in the box and the lettering includes it. Check the spelling in the result; long names can wrap. Q: Is the text always spelled right? A: Nano Banana 2 and GPT Image 2 are reliable for short slogans. If a letter slips, generate again; it costs a few credits. Q: Can I use it as the actual sign? A: Print it on a large sheet, or use it as the layout to copy onto poster board. Q: Can I change the colours? A: Yes, add “school colours navy and gold” in the box. Q: What about a prom sign? A: Replace HOCO with PROM in the prompt; everything else works the same. Q: How much does it cost? A: A few credits per sign. Free credits at sign-up cover several tries. ### AI Melt Effect https://tanvo.ai/video/ai-melt-effect · video · Reactions & effects Your photo stands still for a beat, then the whole figure melts down into a puddle of chrome, wax, honey or ice cream. Six materials, one upload. Looks: Liquid chrome, Candle wax, Ice cream, Honey, Rainbow paint, Melt and re-form. Q: Which photo works best? A: Standing, half or full body, plain background. Busy backgrounds blend into the puddle. Q: How long is the clip? A: Four to ten seconds; six is the default. Q: Does it have sound? A: Yes, a subtle liquid sound is generated with the picture. Q: Can it melt back? A: Pick “Melt and re-form” or type it in the box. Q: How much is it? A: From about 130 credits for six seconds at 480p on Seedance 2 Mini. The table in the generator shows every option. Q: Can I melt an object instead? A: Yes, upload a product or a toy. The prompt treats it the same way. ### Disney Boyfriend Filter https://tanvo.ai/image/disney-boyfriend-filter · image · Styles The TikTok trend: your selfie becomes a hand-drawn animated heroine, with a princely animated boyfriend beside you. Original characters in the classic fairytale style. Looks: Ballroom, Balcony at night, Village market, Lantern boat, Snowy forest, Just me. Q: Is this affiliated with Disney? A: No. Tanvo is not affiliated with Disney, and nothing here draws a character from any film. Every result is an original character of you in a classic hand-drawn fairytale style. Q: Will it still look like me? A: Yes. Face shape, hair colour and features are kept; the rendering style changes. Q: Can I pick a girlfriend instead? A: Yes, use the “Girlfriend instead” chip or type it in the box. Q: Can I describe the boyfriend? A: Yes, hair, outfit, props. One line is enough. Q: How much is it? A: A few credits per image; free credits at sign-up cover several scenes. Q: Can I do a couple photo? A: Yes, upload a photo of you both and say “draw us both as the leads” in the box. ### South Park AI Filter https://tanvo.ai/image/south-park-ai-filter · image · Styles One selfie, one flat paper-cutout you with the deadpan stare, standing at the bus stop in the snow. The TikTok cutout trend, as an original character. Looks: Bus stop, Classroom, On the couch, Building a snowman, With my dog, At the office. Q: Is this affiliated with South Park? A: No. Tanvo is not affiliated with South Park or its makers, and nothing here draws a character from the show. Every result is an original character of you in a construction-paper cutout style. Q: Will it look like me? A: In the way the style allows: your hair colour and shape, your jacket, your glasses if you have them. Q: Can I do my pet? A: Yes, pets come out well in this style. Q: Can I do a group? A: Yes, everyone in the photo becomes a cutout. Q: How much is it? A: A few credits per image; Nano Banana 2 Lite is the cheapest option here. Q: Can I change the background? A: Type it in the box, for example “at a ski lift” or “in my kitchen”. ### Minecraft Parkour Video Generator https://tanvo.ai/video/minecraft-parkour-video-generator · video · Motion & camera A first-person run across floating blocks, made from a line of text: footage nobody else has posted, ready to sit under your clip. Q: Is this real Minecraft footage? A: No. It is generated footage of an original block-style world, so you are not reposting someone else's screen recording. Tanvo is not affiliated with Mojang or Microsoft. Q: How do I make the split-screen? A: Put your talking clip in the top half of a 9:16 canvas and this clip in the bottom half. CapCut, InShot and Premiere all do it with two layers. Q: How long is a clip? A: 5 to 15 seconds per run on the default model. Loop it, or make two runs and join them, for longer videos. Q: Can I change the route? A: Yes. Describe the place and the jumps in the prompt: "through a desert temple, over sand blocks and cactus gaps". Keep the first-person arm line so it still reads as parkour. Q: Does it have sound? A: Footsteps, landings and wind. Mute it under a clip that already has audio. Q: What does it cost? A: The price is shown before you run, about 220 credits for 10 seconds at 480p. Failed runs are refunded. ### Sora 2 Prompt Generator https://tanvo.ai/video/sora-2-prompt-generator · video · Motion & camera The shot-list structure from OpenAI's Sora 2 guide, ready to edit, rendered on Seedance or Veo. Sora 2 itself is no longer available. Q: Can I still use Sora 2? A: No. OpenAI closed the Sora app and its API in 2026. Prompts written for it still work well on Seedance 2.0 and Veo 3.1, which is what this page renders on. Q: Is Tanvo affiliated with OpenAI? A: No. Sora is a trademark of OpenAI. This page uses the publicly documented prompt structure, not any OpenAI service. Q: How many shots fit in 10 seconds? A: Two or three. Give each shot at least three seconds so the camera move has time to read. Q: How do I add dialogue? A: Write it on its own line with who says it: Dialogue: the violinist, softly, "One more song." Seedance 2.0 and 2.5 speak it in sync. Q: Will my old Sora prompts work? A: Mostly. Keep the shot timings within the model's length (15 seconds on Seedance, 8 on Veo) and drop anything about Sora-only features like cameos. Q: What does it cost? A: About 450 credits for 10 seconds at 720p on Seedance 2 Mini, more on the larger models. The price is shown before you run. ### AI Cakeify Effect https://tanvo.ai/video/ai-cakeify-effect · video · Reactions & effects One full-body photo, one clean cut, sponge and buttercream inside. The is-it-cake reveal, with the slice sound. Q: Is it a video or a photo? A: A short video: about two seconds of your photo standing still, then the knife, then the slice lifted out. Q: What photo works best? A: Full body, front-on, standing straight against a plain background. Busy backgrounds and sitting poses make the cut harder to read. Q: Can I choose the flavour? A: Yes. Replace the cake line in the prompt with the layers you want, for example "dark chocolate sponge, ganache and raspberries". Q: Does it work on pets and objects? A: Yes. Upload a photo of the pet or the object instead; the cut and the layers work the same. Q: Does it have sound? A: Yes, a soft slice and a patter of crumbs. Add music in your editor if you like. Q: What does it cost? A: About 130 credits at 480p for 6 seconds on the default model. The price is shown before you run, and failed runs are refunded. ### Raindance AI Trend https://tanvo.ai/video/raindance-ai-trend · video · Duets & trends Start on the photo, hard cut to the music-video shot: the pier, the low sun, you singing straight down the lens. Add the Raindance sound when you post. Looks: Golden-hour pier, Rain on the pier, Waist-deep at sunset, Reception lights, The orange room, Duet with your pet. Q: What is the Raindance trend? A: A short clip that starts on an ordinary photo and cuts on the beat to the same person singing on a sunset pier, posted with the song Raindance. It is a singing trend more than a dance. Q: Does the clip include the song? A: No. Add the Raindance sound in TikTok, Reels or CapCut when you post; that keeps the music licensed through the platform. Q: Will it still look like me? A: Yes. Every scene starts on your exact photo and keeps your face, hair and clothes through the cut. Q: Can I do it with my pet? A: Pick "Duet with your pet" and add a second photo of the animal. That look runs on Seedance 2 Mini, which can take two photos. Q: How long is the clip? A: Eight seconds, the length of the hook. You can go to ten on the default model. Q: What does it cost? A: About 120 credits for 8 seconds at 480p on MiniMax H3. The price is shown before you run, and failed runs are refunded. ### Multiaverso AI Filter https://tanvo.ai/image/multiaverso-ai-filter · image · Face filters One selfie, six versions of you from other universes. The same face every time, so your friends recognise you in all of them. Looks: Glow up, Blonde version, 1970s you, You at 11, 3D cartoon, You at 80. Q: Multiaverso or Multiverso? A: Both spellings are used for the same trend; Multiaverso is the most common hashtag. It is the "other universe versions of you" filter. Q: Can I do it without CapCut? A: Yes. Upload a photo here and download each version; you can still put them together in CapCut or post them as a carousel. Q: Will it still look like me? A: That is the point of the trend, so every prompt starts by keeping your face shape, eyes, nose, smile and skin tone. The age versions keep the same features at another age. Q: Can I write my own universe? A: Yes. Edit the prompt and keep its first line, the one that keeps your face, then describe the new hair, clothes and place. Q: Which model does it use? A: Nano Banana 2, the best of our models at keeping a face through big changes. Nano Banana Pro is sharper; Lite is cheaper. Q: What does it cost? A: About 25 credits per version at 1K. Six versions cost about 150 credits. ### AI Christmas Card Generator https://tanvo.ai/image/ai-christmas-card-generator · image · Photoshoots & family Upload the family photo you have; get the card you meant to take: matching knits, the lit tree, the fireplace and "Merry Christmas" in gold. Q: Can I put the whole family on one card? A: Yes. Upload one photo with everyone in it, or up to three photos of different people, and the card keeps the same number of people. Q: Can I change the greeting? A: Yes. Change the words in quotes in the prompt. Short greetings in quotes come out spelled correctly almost every time; check before you print. Q: Will everyone still look like themselves? A: Yes. Every face, age and hair colour is kept; only the clothes, the room and the light change. Q: What resolution do I need for printing? A: 2K (the default) for a 5×7 card, 4K for anything larger. Q: Can I add the pet? A: Add the pet's photo as a second photo and write "with our dog in a red scarf" into the prompt. Q: What does it cost? A: About 40 credits at 2K and 25 at 1K. The price is shown before you run. ### AI Birthday Photoshoot Generator https://tanvo.ai/image/ai-birthday-photoshoot-generator · image · Photoshoots & family Gold number balloons, a balloon garland, a candlelit cake and studio light, with your face kept exactly. Change the number to your age. Q: Can I put my age on the balloons? A: Yes. Change the number in quotes in the prompt, for example "25" or "40". Two digits come out reliably. Q: Which selfie works best? A: A sharp photo, face front-on and evenly lit, no heavy filter. Your expression carries over, so smile if you want to be smiling. Q: Will it still look like me? A: Yes. The face, hair, skin tone and build are kept; only the outfit, the set and the light change. Q: Can I do a group or couple birthday? A: Upload a photo with both or all of you in it and edit the prompt to "Keep everyone from the photo". Q: Can I print it? A: Yes. Choose 2K or 4K before you run for a sharp print. Q: What does it cost? A: About 25 credits at 1K. The price is shown before you run. ## Articles ### AI arrest filter: walked out in handcuffs, completely unbothered https://tanvo.ai/blog/ai-arrest-filter-guide · Trends · Oct 6, 2026 What the trend is The arrest filter clip shows someone being walked out in handcuffs, flanked by officers, looking far too relaxed about it. The person might be holding a coffee, still dressed for a wedding, or carrying shopping bags. The humor comes from the contrast between the serious situation and the completely unbothered face. The AI Arrest Filter app on Tanvo makes the clip from one photo of you. Why it works It's a visual punchline that needs no setup. Everyone recognizes the "walked out" shot from the news, so dropping a calm, ordinary person into it is instantly funny. It also pairs well with captions about small crimes: eating the last slice, leaving the group chat on read, taking the good parking spot. The important part is that it stays obviously staged comedy. The looks are built around everyday, harmless situations for that reason. The joke works because nobody could mistake it for real. The looks There are five looks, each a different "where were you when they came for you" scenario: - Walked out at night: the classic version, under night lights. - Coffee in hand: still holding a coffee, entirely calm. - Escorted from the wedding: dressed up and led out of a celebration. - Grocery run: mid-errand, shopping in hand. - Karaoke night: led away from the mic. How to make it on Tanvo 1. Open the app. Go to AI Arrest Filter. 2. Pick a look from the list above. 3. Upload your photo. The app asks for one photo: Your photo. 4. Generate. The app runs on MiniMax H3 and makes a 6 second clip at 480p in 16:9. The default costs 90 credits, and if a run fails, the credits are refunded. Tips for a good result Use a calm, front-facing photo. The whole joke is the unbothered expression, so a relaxed face works best. A neutral look or a slight smile reads as "not worried at all." A photo where you're mid-laugh or pulling a face can work against it. Show some of your upper body. The clip needs your hands and arms for the handcuffs and the walk. A photo from the waist up gives the model more to work with than a tight face crop. Keep the light even. A clear, well-lit face holds up better through the motion than a face half in shadow. Pick the look for the caption. Some pairings: - A general "they finally caught me" post: Walked out at night. - Morning-person or work jokes: Coffee in hand. - Wedding season, or the friend who always gives a long speech: Escorted from the wedding. - Snack or shopping crimes: Grocery run. - The friend who never gives up the mic: Karaoke night. Keep it clearly a joke. Use it on yourself, or on friends who are in on it. Write a caption that makes it plain the clip is a filter, and don't post it in a way that could be taken as a real news event about a real person. FAQ What is the AI arrest filter? It's a short video trend where a person is shown being walked out in handcuffs while looking calm and unbothered, played for laughs. On Tanvo you make it with AI Arrest Filter from a single photo. Which model makes it? The app runs on MiniMax H3, a model that makes quick clips from a prompt or a photo. You can read more about it on the MiniMax H3 page. Why do the officers look odd? The officers are generated from scratch around you, and they're background figures rather than the focus of the clip, so they sometimes come out less detailed than your face. If they look wrong in a way that spoils the clip, generate again. Trying a different look also changes the scene around you. How much does it cost? The default, a 6 second clip at 480p in 16:9, is 90 credits. If a run fails, the credits are refunded. See the pricing page for credit packs. Ready to try it? Open AI Arrest Filter. ### AI baby photoshoot: six studio-style shoots from one photo https://tanvo.ai/blog/ai-baby-photoshoot-guide · Trends · Oct 6, 2026 What the trend is Seasonal baby photos are a tradition: the pumpkin patch in autumn, the cake smash on a first birthday, the tiny baby asleep in a basket, the Christmas card. Booking a studio for each one is expensive, and babies rarely cooperate on schedule. The AI Baby Photoshoot app on Tanvo makes studio-style shoots from the one clear photo of your baby you already have on your camera roll. Same face, no studio, no nap schedule. Why it works Parents want the classic shots without the logistics. A single good photo is enough to try every setting, and the results are easy to share with family or use for cards. It's especially useful for milestones you missed or for shoots that would be difficult in real life, like a newborn basket for a baby who is already crawling. The looks There are six looks: - Pumpkin patch: the autumn classic. - First-birthday cake smash: cake, mess and a party setting. - Newborn basket: the sleepy, wrapped-up studio shot. - Christmas: ready for a holiday card. - Little pumpkin costume: your baby dressed as a pumpkin. - Spring meadow: soft light and flowers. How to make it on Tanvo 1. Open the app. Go to AI Baby Photoshoot. 2. Pick a look from the list above. 3. Upload your photo. The app asks for one: Baby's photo. 4. Make it yours (optional). The chips add a detail: Add a name sign, With our dog, Twins or Older sibling too. 5. Generate. The app runs on Nano Banana 2 and makes a 1K square image. The default costs 25 credits, and failed runs are refunded. Nano Banana Pro is also available in the app. Tips for a good result Pick the clearest face you have. The best photo shows your baby's whole face, eyes open, in soft daylight, without a pacifier, a hand or a blanket covering part of it. The model keeps what it can see, so a clear face is the single biggest factor in a likeness. Avoid heavy filters. Beauty filters and strong color edits change skin tone and features. A plain, unedited phone photo works better. Front or slightly to the side. A face turned well away from the camera is harder to keep recognizable in a new pose. Pick the look for the occasion. - Autumn posts and October cards: Pumpkin patch or Little pumpkin costume. - First birthday announcements and invitations: First-birthday cake smash. - A newborn-style keepsake: Newborn basket. - Holiday cards: Christmas. - Spring and Easter: Spring meadow. Use the chips for family shots. With our dog and Older sibling too bring the rest of the household into the picture. Twins is for when there are two babies to include. Add a name sign puts a name in the scene, which is handy for birth announcements. Make a set. At 25 credits per image, it's easy to run all six looks from the same photo and keep a matching set for the year. FAQ Will it look like my baby? It's built to keep your baby's face from the photo you upload. Likeness depends mostly on the photo: a sharp, front-facing, well-lit face gives the best result. If a result doesn't look right, try a clearer photo before trying a different look. Can I add a name or age sign? Yes. Use the Add a name sign chip. Check the spelling in the result before you print or share it, and generate again if any letters come out wrong. What photo works best? A recent, sharp photo of your baby's face in daylight, eyes open, with nothing covering the face and no heavy filters. Phone photos are fine. How much is it? The default, a 1K square image on Nano Banana 2, is 25 credits. Failed runs are refunded. See the pricing page for credit packs. Ready to try it? Open AI Baby Photoshoot. ### AI birthday photoshoot generator: a glam party portrait from one selfie https://tanvo.ai/blog/ai-birthday-photoshoot-generator-guide · Trends · Oct 6, 2026 What the trend is The AI birthday photoshoot is the portrait people post on the morning of a big birthday: a sequin outfit, giant foil number balloons, a balloon arch, a cake with the candles lit and proper studio light. A real shoot like that means a photographer, a backdrop and a lot of balloons. The AI Birthday Photoshoot Generator on Tanvo does it from one selfie. It keeps your face, hair and expression and builds the set around you. The look There's one look, Gold balloons and cake. You stand in a champagne-gold sequin outfit beside large gold foil number balloons, with a pastel pink and cream balloon garland arching behind you. A two-tier white cake with lit candles sits on a small table in front, and a few strands of gold confetti hang in mid-air. The light is soft studio light from the front, a warm rim light and candle glow on your face. How to make it on Tanvo 1. Open the app. Go to AI Birthday Photoshoot Generator. 2. Upload your photo. The app asks for one: Your photo. 3. Set your age. Change the number in quotes in the prompt. 4. Generate. The app runs on Nano Banana 2 and makes a 1K portrait image in 3:4. Tips that make it work Change the number in quotes. The prompt asks for number balloons showing "30", in quotes. Change it to your age and keep the quotes, for example "25" or "40". Two digits come out reliably. Use a sharp, front-facing selfie. Your face should be evenly lit with no heavy filter. A plain background is fine, since it gets replaced. Smile if you want to be smiling. The prompt keeps your expression from the photo, so the face you upload is the face in the shoot. Change the colors. The colors are written into the prompt, so you can swap them: silver and black, or all pink. Some ideas: - Silver and black for a 40th - All pink for a 21st - A first birthday with the baby - A couple's joint birthday - Champagne toast pose - Sash and crown Doing a couple or a group? Upload a photo with both or all of you in it, and edit the start of the prompt to "Keep everyone from the photo". Pick the size for where it's going. 1K is enough for a post. Choose 2K or 4K before you run if you want to print it. Cost and models The default, a 1K image on Nano Banana 2, is about 25 credits. The price is shown before you run, and failed runs are refunded. Nano Banana Pro, Nano Banana 2 Lite and Seedream 5.0 Lite are also available in the app. See the pricing page for credit packs. FAQ Will it still look like me? Yes. Your face, hair, skin tone and build are kept; only the outfit, the set and the light change. Can I print it? Yes. Choose 2K or 4K before you run for a sharp print. Ready to try it? Open AI Birthday Photoshoot Generator. ### AI cakeify effect: a knife goes in, and you were cake all along https://tanvo.ai/blog/ai-cakeify-effect-guide · Trends · Oct 6, 2026 What the trend is Cakeify is the AI version of the old "is it cake?" videos. Something looks completely normal, a knife goes in, and it was cake the whole time. The AI version does it to people: your photo stands still, a chef's knife slices down the middle, and inside are even sponge layers, buttercream and a thin line of jam. The AI Cakeify Effect app on Tanvo makes the clip from one full-body photo, with the soft knife-through-sponge sound included. Why it works The reveal only lands if everything before the cut is exactly you. The face, the clothes and the pose all come straight from the photo, so for two seconds it looks like an ordinary picture. Then the knife goes in. That order, normal first, then the slice, then the wedge lifted out, is built into the app. The look There's one look, Vanilla sponge reveal. It plays out in three beats: - 0-2s: you stand still exactly as in the photo, front-on, on a plain background. - 2-4s: a large chef's knife enters from the top of the frame and slices slowly down the middle, revealing vanilla sponge layers, white buttercream and a thin line of strawberry jam, with a few crumbs dropping. - 4-6s: the knife eases a wedge out toward the camera. The outside still shows your clothes and colors while the cut face shows the layers. How to make it on Tanvo 1. Open the app. Go to AI Cakeify Effect. 2. Upload your photo. The app asks for one: Full-body photo. 3. Pick the cake. Keep the vanilla preset, or change the cake line in the prompt. 4. Generate. The app runs on Seedance 2 Mini and makes a 6 second vertical clip at 480p in 9:16, with sound. Tips that make it work Use a full-body photo. The whole figure should be in frame, standing straight and front-on. The knife cuts down the middle of the figure, so a cropped photo leaves it nothing to cut through. Keep the background plain. A plain wall behind you works best. Busy backgrounds and sitting poses make the cut harder to read. Let the photo come first. The opening two seconds are your photo held still. Don't rewrite that part of the prompt: the stillness is what sells the surprise. Change the flavor, not the order. To swap the cake, replace only the cake line with the layers you want, for example "dark chocolate sponge, ganache and raspberries". Other ideas: - Chocolate and cherry layers - Rainbow sponge - Red velvet with cream cheese - Your dog, cakeified - A product on the desk - A birthday surprise for a friend Try it on pets and objects. Upload a photo of the pet or the object instead. The cut and the layers work the same. Go sharper if you want crumbs. 6 seconds at 480p is enough for the reveal. For sharper crumbs, use 720p on Seedance 2.0. Cost and models The default, 6 seconds at 480p on Seedance 2 Mini, is about 130 credits. The price is shown before you run, and failed runs are refunded. Seedance 2.0, Seedance 2.5 and Kling 3.0 are also available in the app. See the pricing page for credit packs. FAQ Is it a video or a photo? A short video: about two seconds of your photo standing still, then the knife, then the slice lifted out. Does it have sound? Yes, a soft slice and a patter of crumbs. Add music in your editor if you like. Ready to try it? Open AI Cakeify Effect. ### AI Christmas card generator: the family card, from the photo you already have https://tanvo.ai/blog/ai-christmas-card-generator-guide · Trends · Oct 6, 2026 What the trend is Getting everyone into matching sweaters in front of the tree, all looking at the camera at once, is the hardest photo of the year. Someone blinks, the dog leaves, the lights don't show up. The AI Christmas Card Generator on Tanvo starts from a photo you already have, even a casual one, and restages it as the card: matching knits, fairy lights, a lit fireplace, snow at the window and "Merry Christmas" in gold script along the bottom. Why it works Everyone stays themselves. The prompt keeps each face, hair color, age and build, and the same number of people. Only the clothes, the room and the light change. That's why it runs on Nano Banana 2 rather than a lighter model. The look There's one look, Fireside card. Everyone stands close together in matching cream-and-red knitted sweaters beside a tall tree with warm white fairy lights and red and gold baubles. A lit fireplace with knitted stockings sits behind them, snow falls past a window, and warm firelight falls on the faces. A clean cream border frames the photo, with the greeting in gold script centered in the bottom margin. How to make it on Tanvo 1. Open the app. Go to AI Christmas Card Generator. 2. Upload your photo. The app asks for Your photo: one photo with everyone in it, or up to three if the family is spread across photos. 3. Change the greeting if you like. Edit the words in quotes in the prompt. 4. Generate. The app runs on Nano Banana 2 and makes a 2K portrait image in 3:4. 5. Print. Download it and send it to any print shop. Tips that make it work Keep the greeting in quotes. The prompt writes the greeting as "Merry Christmas", in quotes, followed by "spelled exactly like that". To change it, replace only the words inside the quotes, for example "Happy Holidays from the Parks". Short greetings in quotes come out spelled correctly almost every time. Check the spelling before you print anyway. Clear, front-on faces. Every face in the photo should be clear and facing the camera. The card keeps what it can see, so a face turned away or in shadow is harder to keep. Up to three photos. If one photo doesn't have everyone, upload up to three photos of different people. The card keeps the same number of people as your photos. Add the pet. Add the pet's photo as a second photo and write "with our dog in a red scarf" into the prompt. Pick the resolution for the print. 2K, the default, is right for a 5×7 card. Choose 4K for anything larger. Make one for everyone. Some ideas: - The whole family by the fire - Just the two of you - With the dog in a scarf - Grandparents' card - A card from the office team - Baby's first Christmas Cost and models The default, a 2K image on Nano Banana 2, is about 40 credits, and 1K is about 25. The price is shown before you run, and failed runs are refunded. Nano Banana Pro, Nano Banana 2 Lite and GPT Image 2.5 are also available in the app. See the pricing page for credit packs. FAQ Can I put the whole family on one card? Yes. Upload one photo with everyone in it, or up to three photos of different people, and the card keeps the same number of people. Will everyone still look like themselves? Yes. Every face, age and hair color is kept; only the clothes, the room and the light change. Ready to try it? Open AI Christmas Card Generator. ### AI crying filter: the smile that breaks into tears https://tanvo.ai/blog/ai-crying-filter-guide · Trends · Oct 6, 2026 What the trend is The crying filter clip runs in three beats. It opens on a hopeful smile. The smile wavers, and real tears come. Then the person turns and walks away from the camera. It became a popular phone-app template in autumn 2026, often posted with a caption that turns it into a joke, a tribute or a very small heartbreak. The AI Crying Filter app on Tanvo makes the clip from a single photo of you. Why it works It's a whole story in a few seconds: hope, a crack, and an exit. The shift from smile to tears is the hook, because it's a little uncomfortable to watch and very easy to caption. The walk away gives it an ending, so the clip feels finished rather than cut off. It also works for almost any caption, from "me when the group chat goes quiet" to something sincere. The looks There are five looks, each a different way the emotion plays out. - Smile to tears: the classic version, a smile that slowly gives way. - Holding it in: quieter, the person trying not to cry. - Grin to full cry: the big, dramatic version. - Rain on the window: a moodier setting. - Laughing through tears: half funny, half sad, good for jokes. How to make it on Tanvo 1. Open the app. Go to AI Crying Filter. 2. Pick a look from the list above. 3. Upload your photo. The app asks for one photo: Your photo. 4. Generate. The app runs on Seedance 2 Mini and makes a 6 second clip at 480p in 16:9. The default costs 130 credits, and failed runs are refunded. Tips for a good result Start with a neutral or lightly smiling photo. The clip begins on a smile, so a photo where you're relaxed or smiling a little gives the model a natural starting point. A huge open-mouthed laugh is harder to turn into tears convincingly. Make sure your face is clear and close. The tear is a small detail. If your face is tiny in the frame, in shadow or behind sunglasses, the model has less to work with and the tear may not show. A photo from the chest up, in soft even light, is ideal. Include some of your body. The clip ends with you turning and walking away, so a photo that shows your shoulders and some of your upper body helps the turn look natural. Choose the look for the caption. Some pairings: - A sincere or sentimental post: Smile to tears or Holding it in. - An over-the-top joke: Grin to full cry. - A late-night, moody post: Rain on the window. - Something self-aware: Laughing through tears. Keep it obviously a bit. If you post it about another person, make sure the caption makes it clear it's a filter and a joke. FAQ Which app does the AI crying filter? On Tanvo it's AI Crying Filter. It takes one photo and makes the full smile, tears and walk-away clip, using Seedance 2 Mini. Why is the tear missing or the face changing? Most often the photo is the cause. A face that's small in the frame, turned away, in heavy shadow or partly covered gives the model too little to hold on to, so it either skips the tear or redraws the face. Try a closer, front-facing photo with even light. If one look keeps missing the tear, Grin to full cry is the most visible version. Can I change where they walk? The looks control the mood and the setting, so pick the one closest to what you want. Rain on the window gives a different scene from the classic versions. What does it cost? The default, a 6 second clip at 480p in 16:9, is 130 credits. If a run fails, the credits are refunded. See the pricing page for credit packs. Ready to try it? Open AI Crying Filter. ### AI Halloween costume generator: try it on before you buy it https://tanvo.ai/blog/ai-halloween-costume-generator-guide · Trends · Oct 6, 2026 What the trend is Every October, the same question comes up: what are you going as? Trying costumes on your own face before buying anything has become part of the season. Some of the looks people keep returning to are a Y2K vampire shot with harsh camera flash, a cursed forest witch, and couples dressed as matching archetypes. The AI Halloween Costume Generator on Tanvo puts a costume on your own photo, so you can see a witch, vampire or werewolf on yourself before spending anything on the real thing. If you'd rather have the whole mood of a Halloween night, the AI Halloween Photo Trend app reshoots your everyday photo as a flash-lit 1990s Halloween scene. Why it works Costumes are a gamble. Something that looks great on a model in a shop photo can look completely different on you. Seeing it on your own face takes the guessing out of the decision, and it's fun to share. Friends vote, couples plan matching looks, and the images make good October profile pictures in their own right. The costume looks The costume generator has eight looks: - Witch - Y2K vampire - Skeleton face paint - Scarecrow - 1920s flapper - Retro astronaut - Werewolf - Forest fairy How to make it on Tanvo 1. Open the app. Go to AI Halloween Costume Generator. 2. Pick a look from the list above. 3. Upload your photo. The app asks for one photo: Your photo. 4. Generate. The app runs on Nano Banana 2 and makes a 1K square image. The default costs 25 credits, and failed runs are refunded. Generate a few looks side by side. At 25 credits each, trying three or four costumes still costs far less than buying the wrong one. The photo trend version The AI Halloween Photo Trend app takes a different angle. Instead of a costume, it reshoots your photo as a 1990s flash-lit Halloween night with film grain, keeping your face. Its looks are Porch at night, Pumpkin pile, Something in the doorway, Candlelit study, Lantern at the gate and Varsity jacket, pumpkin patch. The chips include With my pet, Couple, Kids too and Office party. It also runs on Nano Banana 2 at 25 credits for the default 1K square, with Nano Banana 2 Lite and Nano Banana Pro available. Tips for a good result Use a clear, front-facing photo. Costumes change your hair, makeup and clothes, so the model needs a good view of your face to keep it yours. Even light and no sunglasses help. Show your shoulders. A photo from the chest up gives the costume room. A tight face crop leaves no space for a cape, a collar or a flapper's beads. Pick the look for the effect you want. - Flash-lit, early-2000s party energy: Y2K vampire. - The cursed forest witch idea: Witch or Forest fairy. - Makeup you could copy at home: Skeleton face paint. - Something classic and elegant: 1920s flapper. - A costume people won't expect: Retro astronaut or Scarecrow. - The full transformation: Werewolf. Plan couple costumes as a pair. For matching archetypes, run each partner's photo through the same look, or pair two looks that belong together, such as Witch and Scarecrow. For one image of you both, the photo trend app has a Couple chip. FAQ How does it work? You upload one photo, pick a costume look, and the app generates an image of you wearing it. It runs on Nano Banana 2, an image model built for generating and editing from photos. Can I do group or couple costumes? The costume generator works from one photo, so run each person separately with matching looks. For a single image with two people, use the Couple chip in AI Halloween Photo Trend. Can I use the result as a reference for a real costume? Yes, that's one of the best uses. Save the image and use it to shop for pieces or to copy the makeup. Skeleton face paint and Y2K vampire are especially easy to recreate at home. What does it cost? Both apps cost 25 credits for the default 1K square image, and failed runs are refunded. See the pricing page for credit packs. Ready to try it? Open AI Halloween Costume Generator. ### AI LARP generator: fake-rich photos, openly staged for laughs https://tanvo.ai/blog/ai-larp-generator-guide · Trends · Oct 6, 2026 What the trend is LARP, borrowed from live-action role play, has come to mean playing a part online. "Larpmaxxing" is the joke version: posting openly staged, fake-rich photos of a lifestyle you clearly don't have. Leaning on a supercar. Stepping off a private jet. Watching the sunset from a penthouse. The point isn't to fool anyone. It's to be so obviously over the top that everyone's in on it. The AI LARP Generator on Tanvo takes one selfie and puts you in the supercar, the jet or the penthouse you don't own. Why it works The humor comes from the gap between the caption and reality. A photo on a yacht deck with a caption like "quiet weekend" lands because your friends know where you actually spent the weekend. It's also a gentle parody of the polished luxury posts everyone has scrolled past, which makes it feel shareable rather than showing off. The looks There are six looks: - Supercar & jet: the full flex, a supercar with a private jet behind it. - Penthouse sunset: a high-floor view at golden hour. - Yacht deck: on the water. - First-class suite: the long-haul seat you didn't pay for. - Rooftop dinner: a city dinner with a view. - Ski chalet: a winter getaway. How to make it on Tanvo 1. Open the app. Go to AI LARP Generator. 2. Pick a look from the list above. 3. Upload your photo. The app asks for one: Your photo. 4. Generate. The app runs on Nano Banana 2 and makes a 1K square image. The default costs 25 credits, and failed runs are refunded. Tips for a good result Use a clear selfie with good light. Your face needs to hold up in a completely new scene, so a well-lit, front-facing photo works best. Daylight from a window is ideal. Pick a confident expression. The joke works best played straight. A relaxed, slightly smug look sells the bit better than a big grin. Avoid busy backgrounds and other people. The app replaces your surroundings, and a clean selfie gives it less to untangle. If friends are in the shot, crop down to just you. Match the look to the caption. Some pairings: - The full "made it" post: Supercar & jet. - "Working from home today": Penthouse sunset or First-class suite. - "Low-key weekend": Yacht deck. - "Just a casual Tuesday dinner": Rooftop dinner. - Winter break jokes: Ski chalet. Make a series. The trend is funnier as a running bit. Generate three or four looks from the same selfie and post them over a week, each with a more understated caption than the last. Keep it obviously a joke. Use it on yourself, write captions that wink at the audience, and don't use the images to mislead anyone about your finances, a sale or a business. The trend only works because it's openly staged. FAQ What is a LARP photo? It's a deliberately fake, over-the-top photo of a lifestyle you don't have, posted as a joke. In the larpmaxxing trend, the more obviously staged the luxury, the better. Will my face look right? The app is built to keep your face from the selfie in a new scene. Likeness depends mostly on the photo: a clear, front-facing face in good light, without sunglasses or a hand in the way, gives the best result. If something looks off, try a sharper selfie. Which model does it use? The app runs on Nano Banana 2, an image model that generates and edits from photos. What does it cost? The default, a 1K square image, is 25 credits. Failed runs are refunded. See the pricing page for credit packs. Ready to try it? Open AI LARP Generator. ### AI melt effect: turn yourself into liquid chrome, wax or ice cream https://tanvo.ai/blog/ai-melt-effect-guide · Trends · Oct 6, 2026 What the trend is The melt effect is a short vertical clip with a satisfying payoff. Your photo stands still for a beat, then the whole figure melts down into a puddle. Depending on the material, you might drip away as shining liquid chrome, soften like a candle, slump like ice cream on a hot day, or pour down as honey. The AI Melt Effect app on Tanvo makes the clip from one photo, with six materials to choose from. Why it works Melting is visually satisfying in the same way slow-motion pours and paint mixing are. The still beat at the start sets up the reveal, so the viewer waits for something to happen, and then it does. It's also very flexible for captions: "me at 3pm on a Monday," "me when they say the meeting could have been an email," or just a clean visual loop. Different materials give it different moods, from sleek to silly. The looks There are six looks: - Liquid chrome: a mirror-like metal melt, sleek and glossy. - Candle wax: a slow, soft drip. - Ice cream: the summer meltdown. - Honey: thick, golden and slow. - Rainbow paint: colorful and playful. - Melt and re-form: you melt down, then come back together. How to make it on Tanvo 1. Open the app. Go to AI Melt Effect. 2. Pick a look from the list above. 3. Upload your photo. The app asks for one: Your photo. 4. Make it yours (optional). The chips change how the melt plays out: Re-form at the end, Melt from the feet up, Slow motion or Rainbow colour. 5. Generate. The app runs on Seedance 2 Mini and makes a 6 second vertical clip at 480p in 9:16. The default costs 130 credits, and failed runs are refunded. Seedance 2 Fast, Seedance 2.0 and Seedance 2.5 are also available in the app. Tips for a good result Use a full or three-quarter body photo. The melt travels down the whole figure into a puddle, so it needs your body in shot, ideally down to the knees or the feet. A tight face crop leaves the model nowhere to pour. Stand against a plain background. A clean wall or an open space makes the melting figure stand out, and leaves floor space for the puddle. Stand still and upright. The clip opens on a still beat, so a natural standing pose with arms relaxed makes that first second look intentional. Pick the material for the mood. - Sleek, fashion or music posts: Liquid chrome. - Cozy, autumn or slow-mood posts: Candle wax or Honey. - Funny "I can't today" posts: Ice cream. - Bright, playful posts: Rainbow paint. - A loop or a twist ending: Melt and re-form. Use the chips for timing. Slow motion stretches the melt for a more dramatic clip. Melt from the feet up reverses the usual direction, which makes a good surprise. Re-form at the end adds a recovery to any material. FAQ Which photo works best? A clear, well-lit photo of you standing, showing most or all of your body, against a simple background. Your face should be visible and not hidden by sunglasses or a hat. How long is the clip? The default is 6 seconds, vertical, at 480p. That's long enough for the still beat, the melt and the puddle. Can it melt back? Yes. Pick the Melt and re-form look, or add the Re-form at the end chip to any other material. How much is it? The default, a 6 second vertical clip at 480p on Seedance 2 Mini, is 130 credits. If a run fails, the credits are refunded. See the pricing page for credit packs. Ready to try it? Open AI Melt Effect. ### AI pumpkin stencil generator: carve a face or a pet https://tanvo.ai/blog/ai-pumpkin-stencil-generator-guide · Trends · Oct 6, 2026 What the trend is Carving a pumpkin of your dog, your partner or your own face is one of the most-shared Halloween projects, and also one of the hardest to get right freehand. The trick is a good stencil: a two-tone design where every dark area is cut away and every light area stays, so light shines through in the right shape. The AI Pumpkin Stencil Generator on Tanvo turns a photo of a face or a pet into a two-tone stencil you can print, tape onto the pumpkin and carve. It can also show you a preview of the finished pumpkin, carved and lit, before you pick up the knife. Why it works A personal pumpkin gets a reaction that a generic jack-o'-lantern doesn't. People recognize their pet's ears or a friend's glasses in the glow. Getting there by hand means turning a photo into simple shapes, which is exactly the part most people find hard. A stencil does that step for you, and the carved preview lets you decide whether a design is worth the effort. The looks There are six looks: - Portrait stencil: a person's face as a two-tone stencil. - Dog stencil: a stencil tuned for a dog's face. - Pet with a moon: your pet set against a moon. - Carved preview, lit: how the portrait looks once carved and lit from inside. - Carved dog, lit: the same preview for a dog. - With a BOO banner: a stencil with a BOO banner added. How to make it on Tanvo 1. Open the app. Go to AI Pumpkin Stencil Generator. 2. Pick a look from the list above. 3. Upload your photo. The app asks for one: Face or pet photo. 4. Make it yours (optional). The chips add details or simplify the design: Add a moon, Add bats, Add a name banner or Simpler, fewer islands. 5. Generate. The app runs on Nano Banana 2 and makes a 1K square image. The default costs 25 credits, and failed runs are refunded. Nano Banana 2 Lite is also available in the app. Tips for a good result Use a photo with strong, simple light. Stencils are built from light and dark. A face or pet lit from one side, with clear shadows, turns into cleaner shapes than a flat, evenly lit photo. Get close. Fill the frame with the face. A small dog in a big garden will become a tiny, fussy stencil. Use a plain background. Busy backgrounds can add stray shapes. A wall, a sofa or the sky behind the subject keeps the stencil clean. Choose "Simpler, fewer islands" for your first pumpkin. An island is a piece of pumpkin skin that would be completely surrounded by cut-out areas, so it falls out when you carve. Fewer islands means a design that holds together, and that's the most common reason a carved pumpkin fails. Check the carved preview. Run Carved preview, lit or Carved dog, lit to see whether the design reads well when it's glowing. If the face isn't recognizable in the preview, it won't be on the porch either. Carving tips that help any stencil: 1. Print the stencil and trim the paper close to the design. 2. Tape it to the flattest side of the pumpkin. 3. Trace the lines with a pin or a poking tool, through the paper. 4. Remove the paper, then cut the smallest, most detailed areas first. FAQ Will the stencil actually be cuttable? It's designed to be, as a two-tone image with cut and keep areas. For an easier carve, add the Simpler, fewer islands chip, and check the result for any small floating pieces before you start cutting. Does it work with pets? Yes. There are pet-specific looks: Dog stencil, Pet with a moon and Carved dog, lit. A close photo of the pet's face, with ears and eyes clearly visible, works best. Can I add text? Yes. With a BOO banner adds a banner to the design, and the Add a name banner chip adds a name. Keep names short, since thin letters are hard to carve. What is the carved preview for? It shows how the stencil will look once carved and lit from inside, so you can judge the design before spending an evening on it. If the preview doesn't read well, try a different photo or the simpler version. Ready to try it? Open AI Pumpkin Stencil Generator. ### Ceiling fan transition: every blade pass, a new outfit https://tanvo.ai/blog/ceiling-fan-transition-guide · Trends · Oct 6, 2026 What the trend is The ceiling fan transition is shot from directly overhead. You lie on the floor or the bed, a ceiling fan spins between you and the camera, and every time a blade sweeps across the frame, your outfit changes. It's a neat trick: the blade hides the cut, so the change looks like it happens by magic. The Ceiling Fan Transition app on Tanvo makes the clip from a full-body photo of you and photos of up to three more outfits. No ceiling fan, no overhead rig and no outfit changes on the floor required. Why it works Outfit transitions are popular because they show several looks in one clip, and the fan gives the format a rhythm. Each blade pass is a beat, and the viewer starts waiting for the next change. The overhead angle also flatters a full outfit, showing it top to bottom in one frame. It works for fashion hauls, "outfits of the week," packing for a trip, or just showing off. The looks There are four looks, each a different room and light: - Bedroom daylight: soft daylight, the classic setting. - Neon night: colored night lighting, good for going-out outfits. - Sunlit floor: lying on the floor in bright sun. - Hotel room: a hotel setting, nice for travel and packing clips. How to make it on Tanvo 1. Open the app. Go to Ceiling Fan Transition. 2. Pick a look from the list above. 3. Upload your photos. The app asks for You, full body, then Outfit 2, Outfit 3 and Outfit 4. The first photo is you in your starting outfit; the others are the outfits you change into. 4. Generate. The app runs on Seedance 2 Mini and makes a 10 second vertical clip at 480p in 9:16. The default costs 220 credits, and failed runs are refunded. Tips for a good result Use a true full-body photo for the first slot. Head to feet, standing straight, arms relaxed, in good light. The clip shows you from above at full length, so cropped feet or a cut-off head have to be guessed. Make the outfit photos clear. Each outfit should be easy to read: the whole garment visible, on a plain background or worn in a simple mirror shot. Flat lays of clothes on a bed work well too, as long as nothing is folded over or hidden. Pick outfits that are different from each other. The trick lands when each change is obvious. A white tee, then a red dress, then a black suit reads clearly. Three similar grey hoodies don't. Match the look to the outfits. - Everyday or casual outfits: Bedroom daylight. - Party and going-out looks: Neon night. - Summer clothes and swimwear: Sunlit floor. - Travel capsules and packing clips: Hotel room. Keep accessories simple. Big hats or bags can confuse where the outfit ends and the room begins. If an accessory matters, make sure it's clearly visible in that outfit's photo. FAQ Do I need a ceiling fan? No. The fan, the room and the overhead angle are all generated. You only upload a photo of yourself and photos of the outfits. How many outfits can I show? Up to four in total: the one you're wearing in the full-body photo, plus Outfit 2, Outfit 3 and Outfit 4. The transitions are not on the blade. Why? The timing of each change is generated, so sometimes a change lands a little before or after the blade passes. Clear, distinct outfit photos help, because the model has an easier job swapping between them. If the timing is off in a way that spoils the clip, generate again. What does it cost? The default, a 10 second vertical clip at 480p on Seedance 2 Mini, is 220 credits. If a run fails, the credits are refunded. See the pricing page for credit packs. Ready to try it? Open Ceiling Fan Transition. ### Hoco sign generator: preview the punny proposal sign first https://tanvo.ai/blog/hoco-sign-generator-guide · Trends · Oct 6, 2026 What the trend is Asking someone to homecoming has become an event in itself. The classic move is a hand-made poster with a pun on it, often paired with a prop: a pizza box that says the question is cheesy, sour candy, a volleyball, a pair of cleats. The sign is the photo everyone posts afterward. The Hoco Sign Generator on Tanvo lets you preview the whole sign, with their name on it, before you buy the poster board or spend an evening with glitter glue. Why it works A good hoco sign is personal and a little corny on purpose. The pun connects to something the other person loves, like their sport, their favorite snack or their band, and the name makes it unmistakably for them. The hard part is design: fitting the words, choosing colors and arranging the props. Seeing a finished version first makes the real one faster to build and better looking. The looks There are six looks, each a different pun and prop: - U + ME = HOCO?: the classic equation sign. - Cheesy pizza box: the pizza-box pun. - Volleyball set: for a volleyball player. - Kick it: a soccer or football style pun. - Glow sticks: bright and party themed. - Sour candy: the candy pun. How to make it on Tanvo 1. Open the app. Go to Hoco Sign Generator. 2. Pick a look from the list above. There's no photo to upload for this one: the sign is made from the look and your details. 3. Make it yours (optional). The chips tailor the theme: Volleyball, Soccer, Football or Band. Add their name so it appears on the sign. 4. Generate. The app runs on Nano Banana 2 and makes a 1K square image. The default costs 25 credits, and failed runs are refunded. GPT Image 2 and Qwen Image 3 are also available in the app. Tips for a good result Keep the words short. Signs work best with a short pun and a name. Long sentences get cramped on a poster and are harder for any image model to letter cleanly. Check the spelling, especially the name. Read every word in the result before you copy it onto a real board. Unusual names are the most likely to come out wrong. If a letter is off, generate again. Try Qwen Image 3 for text-heavy signs. If your sign has more words than usual, Qwen Image 3 is built for readable text and layout, and it's one of the models available in the app. Match the look to the person. - Not sure what they're into: U + ME = HOCO? - Food lover: Cheesy pizza box or Sour candy. - Athletes: Volleyball set or Kick it, plus the matching sport chip. - Band kids: the Band chip. - A night-time ask: Glow sticks. Use the preview as a plan. Note where the big words sit, which colors stand out and where the props go. Sketch the layout lightly in pencil on the real board before you paint or glue anything. FAQ Can it put their actual name on the sign? Yes. Add the name when you make the sign and it's written into the design. Short names come out cleanest, so if they go by a nickname, consider using it. Is the text always spelled right? Not always. Image models can still slip on lettering, especially with long or unusual words. Check every word, and generate again if something is off. Qwen Image 3, available in the app, is built for readable text. Can I use it as the actual sign? You can print it, but most people use it as a design to copy onto a real poster board, which shows the effort that makes a hoco ask special. A 1K image is fine for planning and for sharing on a phone. What about a prom sign? The looks are written for homecoming, but the same approach works: keep it short, make it personal, and check the spelling. Ready to try it? Open Hoco Sign Generator. ### Hotel lobby AI video: two people, one mic, one orange booth https://tanvo.ai/blog/hotel-lobby-ai-video-guide · Trends · Oct 6, 2026 What the trend is The hotel lobby video is simple to describe: two people squeezed into a small orange booth, sharing one hanging microphone, trading lines at the camera. One leans in and delivers a verse, the other answers, and the joke is in the pairing. Best friends, a couple, a parent and child, two coworkers, or you and someone you've never met in your life. The Hotel Lobby AI Video app on Tanvo makes the whole clip from two photos. You don't need a booth, a mic or the other person in the same room. Why it works The format has a lot going for it. The setting is instantly recognizable, so viewers know what's coming in the first second. The single shared mic forces the two people close together, which makes any pairing funny or sweet. And because the structure is always the same, the people are the whole story. That's why it's a good fit for AI: you supply the faces, and the format supplies the rest. The looks There are six looks. Each keeps the booth and the shared mic and changes the styling around them. - Classic booth: the original orange booth and hanging mic. - Wedding day: the booth as a wedding moment, good for couples and anniversaries. - Rap battle: more attitude, more back-and-forth. - Matching suits: both people dressed alike, which suits duos and coworkers. - Halloween: the booth in a seasonal mood. - Retro 70s: warmer, older styling. How to make it on Tanvo 1. Open the app. Go to Hotel Lobby AI Video. 2. Pick a look from the list above. 3. Upload two photos. The app asks for Person 1 (left) and Person 2 (right). Put each photo in the slot for the side you want that person on. 4. Make it yours (optional). Add a detail with the chips: Matching suits, Wedding outfits, Streetwear or Birthday. These change the styling without changing the format. 5. Generate. The app runs on Seedance 2.5 and makes a 10 second vertical clip at 480p in 9:16, with sound, in about three minutes. The default costs 600 credits. If a run fails, the credits are refunded. Tips for a good result Choose clear, front-facing photos. Each person should be looking roughly toward the camera, with their face well lit and not covered by sunglasses, a hat brim or a hand. A photo taken at arm's length works better than a group shot cropped down to one face. Match the photos if you can. Two photos with similar light and similar framing, both from the chest up for example, sit together more naturally in the booth than one bright selfie and one dim party photo. Use one person per photo. If the photo you have includes other people, crop it so only the person you want is in frame. The app needs to know whose face goes where. Pick the look for the pairing. Some quick suggestions: - Couple or anniversary: Wedding day, or Classic booth with the Wedding outfits chip. - Friends who roast each other: Rap battle. - Coworkers, siblings or a work duo: Matching suits. - An October post: Halloween. - A parent and an older relative: Retro 70s. Don't fight the format. The charm is the orange booth and the single mic. Use the chips to add a twist, and leave the booth itself alone. FAQ Which app makes the hotel lobby AI video? On Tanvo it's Hotel Lobby AI Video. It takes two photos, one per side of the booth, and generates the full clip with sound. By default it runs on Seedance 2.5, and Seedance 2.0, Seedance 2 Fast and Seedance 2 Mini are also available in the app. Why does one of the faces not look like the photo? Usually the photo is the cause: a face turned away, in shadow, small in the frame or partly hidden. Swap in a clearer, closer photo of that person and generate again. Similar lighting in both photos also helps the two faces hold up equally. Does it use the original footage or audio? No. The clip is generated fresh from your two photos and the look you picked, with its own sound. Nothing is taken from any existing video. What does it cost? The default, a 10 second vertical clip at 480p on Seedance 2.5, is 600 credits. The price for your settings is shown before you generate, and failed runs are refunded. See the pricing page for credit packs. Ready to try it? Open Hotel Lobby AI Video. ### Minecraft parkour video generator: original footage for the bottom of your split-screen https://tanvo.ai/blog/minecraft-parkour-video-generator-guide · Trends · Oct 6, 2026 What the trend is Split-screen posts put something to listen to on top, like a story, a podcast moment or a talking clip, and a parkour run underneath. The run gives the eye something to follow while the audio plays. The catch is that most people reuse the same few screen recordings from the same few channels, and viewers notice. The Minecraft Parkour Video Generator on Tanvo makes the bottom half fresh instead: a first-person run across floating blocks, made from a line of text. Footage nobody else has posted. One note before you start: Tanvo is not affiliated with Mojang or Microsoft. The clips are original footage of a generic block-style world, not recordings of the game. How to make it on Tanvo 1. Open the app. Go to Minecraft Parkour Video Generator. 2. Keep or edit the run. The preset prompt describes a sky-island route. Change the setting or the jumps in plain words. 3. Generate. There's nothing to upload, it's text only. By default the app runs on Seedance 2 Mini and makes a 10 second vertical clip at 480p in 9:16, with sound. 4. Stack it. In your editor, put your clip in the top half of a 9:16 canvas and this one in the bottom half. CapCut, InShot and Premiere all do it with two layers. Mute the run if your top clip has sound, and loop it to length. The look There's one look, Sky islands run. It sprints along a one-block-wide stone path floating above a sea of clouds, chains three jumps up a staircase of floating oak and cobblestone blocks, then makes a long jump to a distant island of grass and oak trees. The player's square blocky arm swings at the lower right of the frame the whole way. For variety, swap the setting in the prompt. Some routes to try: - Desert temple route - Snowy mountain jumps - Night run with torches - Lava cave bridges - Ocean monument path - Cherry-blossom islands Tips that make it work Keep the timings. The preset is split into three beats, 0-3s, 3-6s and 6-10s, each with its own action: a sprint and a gap jump, then a turn and a jump chain, then the long jump. When you change the route, change what happens inside each beat and keep the times. That gives the run a steady rhythm. Keep the arm line. The line about the blocky arm at the lower right is what makes the clip read as first-person parkour. Change the world around it, not the arm. Describe jumps in plain words. Something like "through a desert temple, over sand blocks and cactus gaps" works better than a pile of adjectives. Name the blocks, the gaps and the turns. Leave the frame to the world. The preset asks for a clean gameplay look with the whole frame given to the world and the arm. That's what you want under a talking clip, since anything busy competes with the top half. Pick the length for the screen. 10 seconds at 480p is the cheapest loopable length. If the run fills a whole screen, go to 15 seconds or 720p. For longer videos, loop the clip, or make two runs and join them. Cost and models The default, 10 seconds at 480p with sound on Seedance 2 Mini, is about 220 credits. The price is shown before you run, and failed runs are refunded. Seedance 2.0, Seedance 2.5, Veo 3.1 and Kling 3.0 are also available in the app. See the pricing page for credit packs. FAQ Is this real Minecraft footage? No. It's generated footage of an original block-style world, so you're not reposting someone else's screen recording. Tanvo is not affiliated with Mojang or Microsoft. How long is a clip? 5 to 15 seconds per run on the default model. Does it have sound? Yes: footsteps, landings and wind. Mute it under a clip that already has audio. Ready to try it? Open Minecraft Parkour Video Generator. ### Multiaverso AI filter: six versions of you from other universes https://tanvo.ai/blog/multiaverso-ai-filter-guide · Trends · Oct 6, 2026 What the trend is Multiaverso, also written Multiverso, is the filter trend that asks what you would look like in another universe. It started in Latin American and Brazilian CapCut edits and spread as carousels: the original selfie first, then the glow-up, the blonde version, the 1970s version, the kid version, the cartoon version and the very old version. The Multiaverso AI Filter app on Tanvo has all six universes written. You don't need CapCut: upload one selfie, pick a version, and download each one. You can still put them together in CapCut afterwards, or post them as a carousel. Why it works The whole trend depends on friends recognizing you in every slide. So each version keeps your face shape, eyes, nose, smile, skin tone and any freckles, and changes everything else: hair, clothes, era, place. The age versions keep the same features at another age. The looks There are six universes: - Glow up: a professional glow-up, glossy styled hair, polished grooming and a tailored black outfit, shot as a magazine beauty portrait. - Blonde version: long honey-blonde waves at a sunny café table on a whitewashed seaside street. - 1970s you: a real 1970s family-album photo, feathered hair, a corduroy jacket and a wood-panelled living room, on warm faded film. - You at 11: you at about eleven, in a classic school portrait with a proud gap-toothed grin. - 3D cartoon: you as an original 3D animated film character, waving from a sunny town square. - You at 80: you at about eighty, on a park bench in autumn with reading glasses on a cord. How to make it on Tanvo 1. Open the app. Go to Multiaverso AI Filter. 2. Pick a universe from the list above. 3. Upload your photo. The app asks for one: Your photo. 4. Generate. The app runs on Nano Banana 2 and makes a 1K portrait image in 3:4. 5. Make the set. Run the other five with the same photo. Tips that make it work Use one clear selfie for all six. Head and shoulders, face clear and evenly lit. Glasses are fine; heavy filters are not, because they change the very features the set needs to keep. Same photo, every universe. Using one photo for all six is what makes the carousel read as one person. Swapping photos between versions weakens the effect. Post in order. Put the original selfie first, then the six versions. The original is the reference that makes every other slide land. Write your own universe. Edit the prompt, keep its first line, the one that keeps your face, then describe the new hair, clothes and place. Try it on other people. Some ideas: your parents at 11, the couple version, your best friend at 80, or the office team in the 70s. Cost and models Each version is about 25 credits at 1K, so all six cost about 150 credits, and failed runs are refunded. Nano Banana 2 is the default because it's the best of the app's models at keeping a face through big changes. Nano Banana Pro is sharper, and Nano Banana 2 Lite is cheaper. Seedream 5.0 Lite is also available in the app. See the pricing page for credit packs. FAQ Multiaverso or Multiverso? Both spellings are used for the same trend; Multiaverso is the most common hashtag. It's the "other universe versions of you" filter. Will it still look like me? That's the point of the trend, so every prompt starts by keeping your face shape, eyes, nose, smile and skin tone. Ready to try it? Open Multiaverso AI Filter. ### Pata Chalega AI video: from your photo to the driver's seat https://tanvo.ai/blog/pata-chalega-ai-video-guide · Trends · Oct 6, 2026 What the trend is Pata Chalega is an Indian car trend with a simple shape. The clip starts on an ordinary photo for a beat, then cuts to the same person driving a dark premium car, smiling and waving through the windshield. It's a playful flex: the before is everyday, the after is the big reveal. The Pata Chalega AI Video app on Tanvo makes the drive from one photo of you. You end up behind the wheel of a black luxury sedan, smiling and waving at the camera. Why it works The cut is the whole joke. A normal photo, then suddenly a confident driver in a dark car: it's a small surprise that viewers get instantly. It works as a celebration (a new job, results day, a birthday), as a joke about big dreams, or simply as a fun way to show a good photo. The windshield framing keeps the focus on your face and your wave, which makes the reveal feel personal. The looks There are six looks, each a different drive: - Sunny drive: bright daylight, the classic version. - Night drive: city lights and a darker mood. - Aviators on the highway: sunglasses and open road. - Rain and wipers: rain on the glass, wipers going. - Village road: a countryside route. - Peace sign: a peace sign instead of a wave. How to make it on Tanvo 1. Open the app. Go to Pata Chalega AI Video. 2. Pick a look from the list above. 3. Upload your photo. The app asks for one photo: Your photo. 4. Generate. The app runs on MiniMax H3 and makes a 6 second clip at 480p in 16:9. The default costs 90 credits, and if a run fails, the credits are refunded. Tips for a good result Use a clear, front-facing photo. In the clip you're seen through the windshield, looking at the camera. A photo where you face the camera, with your eyes visible and your face evenly lit, translates best. Head and shoulders is enough. The driver's seat only shows you from the chest up, so a full-body photo isn't needed. A closer photo gives the model more detail on your face. Avoid sunglasses in your photo, unless you choose Aviators on the highway, where the look adds them for you. In the other looks, hidden eyes make it harder to keep your face recognizable. Pick a smile. The clip ends with you smiling and waving. A photo where you're already smiling, or at least relaxed, makes that ending look natural. Choose the look for the moment. - A celebration post: Sunny drive or Peace sign. - A late-night or city mood: Night drive. - The "main character" version: Aviators on the highway. - Something more cinematic: Rain and wipers. - A going-home or hometown post: Village road. Plan the cut. The trend opens on your photo before cutting to the drive. Post your original photo as the first beat in your editor, then cut to the clip on the beat of your audio. FAQ Do I need a photo of a car? No. The car, the road and the setting are all generated. You only upload a photo of yourself. Which model is used? The app runs on MiniMax H3, which makes quick clips from a prompt or a photo. You can read more on the MiniMax H3 page. Can I use a photo where I'm wearing sunglasses? You can, but a photo with your eyes visible keeps your face more recognizable. If you want the sunglasses in the clip, upload a clear photo without them and pick Aviators on the highway. What does it cost? The default, a 6 second clip at 480p in 16:9, is 90 credits. If a run fails, the credits are refunded. See the pricing page for credit packs. Ready to try it? Open Pata Chalega AI Video. ### Raindance AI trend: from an ordinary photo to singing on a golden-hour pier https://tanvo.ai/blog/raindance-ai-trend-guide · Trends · Oct 6, 2026 What the trend is The Raindance trend opens on a completely ordinary photo: a sidewalk selfie, a desk, a kitchen. Then it cuts on the beat to a music-video shot of the same person singing softly on a wooden pier at golden hour. People post it with the Raindance sound, and the joke is the gap between the two shots. It's a singing trend more than a dance. The Raindance AI Trend app on Tanvo does the cut for you from one photo. The clip comes back without the song. Add the Raindance sound from TikTok's library when you post, or in Reels or CapCut, which keeps the music licensed through the platform. The looks There are six scenes, each with its own light and camera move: - Golden-hour pier: sitting on the edge of a weathered pier, the low sun behind you, a slow push-in as you sing to the lens. - Rain on the pier: the far end of a long pier at blue hour, light rain falling through the glow of an old lamp post. - Waist-deep at sunset: standing in a calm, shining sea, trailing a hand across the water. - Reception lights: an outdoor evening wedding reception under strings of warm bulbs, the guests a soft blur behind you. - The orange room: a bare room washed in deep orange sunset light from a tall window. - Duet with your pet: you and your pet on warm sand by the sea at golden hour. How to make it on Tanvo 1. Open the app. Go to Raindance AI Trend. 2. Pick a scene from the list above. 3. Upload your photo. The app asks for Your photo. For the pet duet, add a second photo in Your pet (optional). 4. Generate. The app runs on MiniMax H3 and makes an 8 second vertical clip at 480p in 9:16. The Duet with your pet look runs on Seedance 2 Mini instead, which can take two photos, and comes with sound. 5. Add the sound and post. Put the Raindance sound on it in TikTok or CapCut. Eight seconds fits the hook. Tips that make it work Use an ordinary photo. Every scene starts on your exact photo, held still for the first second, then hard cuts to the music-video shot. The more ordinary the photo, the better the cut lands. A desk selfie beats a polished portrait here. Keep the face clear. A selfie or a waist-up photo with your face clear works best. Your face, hair and clothes are kept through the cut, so the clearer they are in the photo, the more the second shot looks like you. Line the cut up with the beat. The cut happens at the one second mark. When you add the sound, slide it so the beat drops on the cut. Pick the scene for the mood. - The classic version: Golden-hour pier. - A slower, sadder post: Rain on the pier. - Something dreamy and romantic: Waist-deep at sunset. - Wedding season: Reception lights. - A quieter, more intimate clip: The orange room. - The funny one: Duet with your pet. Make a set. Some ideas: from your desk to the pier, grandma's turn, your whole friend group with one clip each, the office kitchen edition, or a couples version. Cost and models The default, 8 seconds at 480p on MiniMax H3, is about 120 credits. The price is shown before you run, and failed runs are refunded. You can go to ten seconds on the default model. Seedance 2 Mini, Seedance 2.0 and Kling 3.0 are also available in the app. See the pricing page for credit packs. FAQ Does the clip include the song? No. Add the Raindance sound in TikTok, Reels or CapCut when you post. That keeps the music licensed through the platform. Will it still look like me? Yes. Every scene starts on your exact photo and keeps your face, hair and clothes through the cut. Ready to try it? Open Raindance AI Trend. ### Sora 2 prompt generator: write the shot list, render it on Seedance or Veo https://tanvo.ai/blog/sora-2-prompt-generator-guide · Trends · Oct 6, 2026 What the trend is Sora 2 taught a lot of people a better way to write video prompts. Instead of a pile of adjectives, you write a short brief that reads like a shot list. Each shot gets its own time, one camera move and concrete things to see, and sound gets its own line. OpenAI closed the Sora app and its API in 2026, so Sora 2 itself is no longer available. The structure still works on the models that are running. The Sora 2 Prompt Generator on Tanvo starts you with a brief already written in that shape and renders it on Seedance or Veo. Tanvo is not affiliated with OpenAI; Sora is OpenAI's trademark. How to make it on Tanvo 1. Open the app. Go to Sora 2 Prompt Generator. 2. Edit the brief. Replace the scene and the three shots with yours, and keep the shape. 3. Add a start frame (optional). The app takes one optional photo, Start frame (optional), to open the clip. With a photo in place, the brief only needs to describe the motion. 4. Generate. By default the app runs on Seedance 2 Mini and makes a 10 second clip at 720p in 16:9, with sound. 5. Adjust. If one shot comes out wrong, change only that shot's line and run again. The preset brief There's one look, Three-shot brief. Its scene is a narrow market street at night after rain, red paper lanterns reflected in wet cobblestones, a street violinist playing under a shop awning. The rest is laid out line by line: - Look: 35mm lens, 16:9, warm lantern highlights against cool blue shadows, soft film grain, shallow depth of field. - Shot 1 (0-3s): wide and locked off, the violinist small in frame. - Shot 2 (3-7s): a slow dolly in at chest height to a medium shot. - Shot 3 (7-10s): a close-up on her face as she smiles at a child dropping a coin into the violin case. - Sound: a slow solo violin melody, rain dripping, quiet street chatter. Tips that make it work Make the timings add up. The shot times should add up to the clip length. Two or three shots fit in 10 seconds. Give each shot at least three seconds so its camera move has time to read. One camera move per shot. Locked off, a slow dolly, a close-up. Use real nouns and verbs. Steam rising from a food stall, raindrops falling from the edge of the awning. Concrete things to see work better than mood words. Give sound its own line. Put dialogue on its own line too, with who says it: Dialogue: the violinist, softly, "One more song." Seedance 2.0 and 2.5 speak it in sync. Keep the consistency line. The preset ends by asking for the same lens, grade and light across all three shots so they read as one scene. Keep that line in your own briefs. Bring your old Sora prompts. They mostly work. Keep the shot timings within the model's length, 15 seconds on Seedance and 8 on Veo, and drop anything about Sora-only features like cameos. Some briefs to try: a product reveal in three shots, two people with one line of dialogue each, or a chase down a staircase. Which model, and the cost - Seedance 2 Mini to draft cheaply. The default, 10 seconds at 720p, is about 450 credits; failed runs are refunded. - Seedance 2.0 or Seedance 2.5 for the final, with sound and dialogue, up to 15 seconds. - Veo 3.1 for 8 second shots with a film look. - Kling 3.0 is also available in the app. The larger models cost more. The price for your settings is shown before you run. See the pricing page for credit packs. FAQ Can I still use Sora 2? No. OpenAI closed the Sora app and its API in 2026. Prompts written for it still work well on Seedance 2.0 and Veo 3.1, which is what this app renders on. Is Tanvo affiliated with OpenAI? No. Sora is a trademark of OpenAI. The app uses the publicly documented prompt structure, not any OpenAI service. Ready to try it? Open Sora 2 Prompt Generator. ### Which image models can spell? https://tanvo.ai/blog/poster-text · Reviews · Oct 6, 2026 Text inside images used to be the giveaway that a picture was AI-made: a "CONCRET" here, a "Satruday" there. So we set one poster brief on six image models available on Tanvo, ran each once at its default size, and checked every letter. The brief Every model got exactly the same prompt: - A concert poster in a two-colour risograph style, teal and orange on cream paper, with a stylised apple tree at night. - Exactly three lines of text, spelled exactly as written: the title MIDNIGHT ORCHARD at the top, Live at the Glasshouse · Saturday 14 November in the middle, and Doors 7:30 PM — Tickets at the gate at the bottom. - No other words, letters or logos. That is fifteen words, a middle dot, an em dash, a time and a date: enough places for a model to slip. Each model ran once, with no retries and no cherry-picking. The result: zero typos All six models spelled all fifteen words correctly. Every word, the date, "7:30 PM", the middle dot and the em dash came out as written, and none of them added stray words or fake logos. A year ago this test would have produced a pile of near-misses; for short, clearly quoted text it is now a solved problem on current models. What still separates them is how closely they follow the rest of the brief. Where they differed Nano Banana 2 set clean, even type and kept to two colours, but moved the middle line down next to the bottom line instead of leaving it in the middle. Nano Banana Pro was the only model that put each line where the brief asked: title at the top, venue and date in the middle, doors and tickets at the bottom. It also had the most convincing print texture. GPT Image 2 chose a heavy vintage display face and a detailed moonlit illustration. The text is perfect; the middle line sits just above the bottom one. GPT Image 2.5 produced a very similar layout with a wider title and the two lower lines grouped together under the artwork. Seedream 5.0 Lite went the other way and placed the venue line directly under the title, leaving the tree to fill the rest of the page. Its type is crisp and modern. Qwen Image 3 kept the most faithful two-colour risograph look, but dropped the night: the tree stands on a plain cream background with no sky. What to take from it - Spelling is no longer the deciding factor for posters, menus, signs and covers with a few short lines. Put the exact text in quotes and say "spelled exactly as written". - Layout is. If the position of each line matters, say so plainly ("in the middle, between the title and the illustration") and expect to re-run once. Nano Banana Pro followed placement best in this test. - Style instructions are followed unevenly. Night, two colours and print texture were each dropped by at least one model. Name the one or two style points that matter most and put them first. - Price differences are small for a single poster. At default settings these models cost between 10 and 50 credits a run on Tanvo, so trying two models side by side is usually cheaper than fixing a layout by hand. This was one brief and one run per model, so treat it as a snapshot rather than a benchmark. If your design has longer paragraphs of text, a different language or very small type, test it the same way: same prompt, several models, check every letter. You can run the same brief yourself in the image generator. ### Zombie AI video and zombie filter: an apocalypse love story https://tanvo.ai/blog/zombie-ai-video-guide · Trends · Oct 6, 2026 What the trend is The zombie trend clip is a tiny apocalypse love story. One person has turned. The other is holding on, gun shaking, unable to do it. Then the clip cuts to a flashback of the two of them on their happiest day. It's dramatic, a little silly and surprisingly touching, which is why couples, best friends and pet owners keep making their own. Tanvo has two apps for it. Zombie AI Video makes the full clip from two photos. AI Zombie Filter turns one photo into a still image with zombie makeup on your own face, from freshly bitten to long gone. Why it works The format borrows the emotional beats of a big zombie film and squeezes them into ten seconds: threat, hesitation, memory. The flashback is the twist that makes people stop scrolling. It also works with almost any pair. Partners, siblings, roommates, and even a person and their dog all fit the story. The video looks Zombie AI Video has six looks: - Field at dusk: the classic open-field standoff at sunset. - Kitchen table: the scene at home, across the kitchen. - They open their arms: a softer ending, with a hug instead of a standoff. - Rainy car park: rain, concrete and headlights. - Pet version: for when the one who turned is your pet. - Last bus: the scene at a lonely bus stop. How to make the video on Tanvo 1. Open the app. Go to Zombie AI Video. 2. Pick a look from the list above. 3. Upload two photos. The app asks for The one holding on and The one who turned. 4. Make it yours (optional). The chips add a setting or a twist: Our kitchen, A field at dusk, A rainy car park or They open their arms instead. 5. Generate. The app runs on Seedance 2 Mini and makes a 10 second vertical clip at 480p in 9:16, with sound. The default costs 220 credits. Seedance 2 Fast, Seedance 2.0 and Seedance 2.5 are also available in the app. Failed runs are refunded. The still version: AI Zombie Filter If you want a photo instead of a clip, or a profile picture for October, use AI Zombie Filter. It puts zombie makeup on your own face (grey skin, clouded eyes, veins up the neck) while keeping you recognizable. Its six looks are Just bitten, Classic walker, Party makeup, Zombie bride, Long gone and Night selfie, and the chips include Keep my glasses, Office background, Rainy street and Torn wedding suit. It runs on Nano Banana 2 and costs 25 credits at the default 1K square size. The two apps go well together. Make a zombie still of the person who turns, then use it as a teaser before posting the full clip. Tips for a good result Use one clear face per photo. Each photo should show one person, facing roughly toward the camera, in even light. Crop group photos down to the person you want. Match the roles to the photos. The first slot is the survivor and the second is the one who turned. Swapping them by mistake is the most common slip. Choose photos with some warmth. The flashback is the emotional payoff, so smiling photos make it land harder. Pick the look for your pair. - Couples: Field at dusk or Kitchen table. - A happier ending: They open their arms, or the chip They open their arms instead. - Moody and cinematic: Rainy car park or Last bus. - You and your dog or cat: Pet version. For the zombie filter, start with a neutral face. The makeup reads best on a calm expression in good light. Pick Just bitten for something subtle and Long gone for the full effect. FAQ Does it keep both faces? Yes, that's what the two photo slots are for. Faces hold up best when each photo is clear, front-facing and well lit. If one face drifts, swap in a closer photo of that person. Can the second photo be a pet? Yes. Choose the Pet version look and upload your pet's photo as The one who turned. A clear photo of the pet's face works best. Can I make a video from the zombie filter? AI Zombie Filter makes a still image. For a clip with motion and sound, use Zombie AI Video. How much does it cost? The video is 220 credits at the default of 480p, 10 seconds, vertical. The zombie filter still is 25 credits at 1K. Failed runs are refunded. See the pricing page for credit packs. Ready to try it? Open Zombie AI Video, or start with the still in AI Zombie Filter. ### Image to video: a setup that works the first time https://tanvo.ai/blog/image-to-video-guide · Guides · Sep 29, 2026 Turning a still into a clip looks simple: upload a photo, type what should happen, press generate. Most first attempts still come back with a face that slowly turns into someone else, hands that melt, or a camera that wanders off. Almost all of that is decided before you press the button. This guide covers the three choices that matter most: the starting frame, the amount of motion, and what you tell the model to keep. Start from the right frame The model treats your photo as frame one. Everything it invents afterward has to grow out of that frame, so a weak start gives a weak clip. A good starting frame has: - A clear, well-lit face. Front or three-quarter view, eyes visible, no sunglasses, no heavy shadow across half the face. The model can only keep what it can see. - The whole subject in shot. If the clip ends with someone walking away, their legs need to exist in the photo. Cropped feet or a cut-off arm get filled in by guesswork. - Room to move. Leave some space in the direction of motion. A person pressed against the right edge has nowhere to walk. - The right shape. Pick a photo that already matches your output, vertical for 9:16 and wide for 16:9. Heavy cropping to fit loses detail. - One main subject. Group shots work, but every extra face is another face that can drift. If you don't have a good frame, make one. Generating a still first with an image model such as Nano Banana 2 costs far less than a video run, and you can fix the light, the outfit and the framing before any motion is involved. Ask for less motion than you think The most common mistake is asking for too much at once. "She laughs, spins around, the camera flies over the city and it starts to snow" gives the model four separate jobs in a few seconds, and it will do all of them badly. Write motion as ordered beats, one after another, and keep it to two or three for a short clip: 1. "She looks into the camera and smiles." 2. "She glances down, then turns her head to the left." 3. "She walks out of frame to the left." Then add one camera move per shot, written separately: "Slow push-in" or "Camera stays locked off." If you want a whip-pan and an orbit, those are two shots. A useful rule: a 5 to 6 second clip holds about two beats comfortably, and a 10 second clip about three or four. If your prompt reads like a paragraph of plot, split it into several clips. Say what must stay the same Models are good at following instructions about change and less reliable about what should stay fixed, unless you say so. Name the things that must not move: - "Her face stays unchanged throughout." - "Same hair, same red jacket, same earrings." - "The background stays the same street." If you only remember one phrase, make it "face stays unchanged". Putting it at the end of the prompt, after the beats, is a simple habit that keeps the model focused on the person you gave it. It also helps to avoid wording that invites change. "She transforms," "becomes," or "morphs" tell the model it's allowed to rebuild the person. If you want an outfit change or a transition, describe exactly what changes and repeat what doesn't. A full example Here is a prompt for a 6 second vertical clip from a waist-up photo of someone outside a café: "She holds a coffee cup and looks at the camera. She smiles and lifts the cup slightly, as if saying cheers. She turns and walks to the right, out of frame. Camera locked off at chest height. Soft afternoon light. Her face, hair and green coat stay unchanged throughout." That's three beats, one camera instruction, one lighting note and one line about what stays the same. It's short, and that's on purpose. Draft cheap, then commit Your first run is a test, so run it on something inexpensive. At the time of writing, Tanvo Video v1 makes 480p drafts at 20 credits for 5 seconds, and Seedance 2 Mini starts at 130 credits for 480p, 6 seconds, with sound. Use the draft to check timing and whether the face holds. Once the motion is right, run the same photo and prompt on the model you want for the final, at the resolution you need. If a run fails, the credits come back. Quick checklist - Clear face, whole subject, matching aspect ratio. - Two or three beats, in order. - One camera move. - A closing line naming what stays the same. - A cheap draft before the expensive run. You can try this on any model from the video page. ### Picking a video model by job, not by hype https://tanvo.ai/blog/pick-a-video-model · Comparisons · Sep 23, 2026 There are a lot of video models on Tanvo, and the newest one is not always the right one. A better question than "which is best?" is "what is this clip for?" A rough draft, a talking shot, a product close-up and a 30 second take have different needs, and each has a sensible place to start. The prices below are each model's default setting at the time of writing. Defaults use different resolutions and lengths, so don't compare the numbers directly. The price for your exact settings is always shown before you generate, and the pricing page explains how credits work. Drafts and tests When you're still working out the motion, the camera move or the timing, you want cheap and fast. - Tanvo Video v1 is the in-house engine: 480p drafts from a prompt or a photo, 5 or 8 seconds, at 20 credits for the 5 second default. This is the place to check whether an idea works at all. - Seedance 2 Mini is the cheapest Seedance tier. It runs at 480p or 720p, 4 to 15 seconds, with sound included, starting at 130 credits for 480p and 6 seconds. It's useful when you need to hear the draft as well as see it. - MiniMax H3 makes quick clips from a prompt or a photo, up to 10 seconds, from 480p up to 1080p. The default 480p, 6 second run is 90 credits. A good habit: draft on one of these, and only move up once the shot is locked. Clips with sound and dialogue If the clip needs voices, ambient sound or music generated with the picture, pick a model that writes audio in the same pass. - Veo 3.1 from Google makes cinematic video with sound and comes in Lite, Fast and Quality tiers. The default, Lite at 720p for 6 seconds, is 115 credits. Lengths are 4, 6 or 8 seconds, so it suits short scenes. - Seedance 2.0 makes video from text or images with native audio, 480p to 1080p, 4 to 15 seconds. The default 720p, 6 second run is 480 credits. - Seedance 1.5 Pro is a mid-weight option with optional audio, 480p to 1080p, up to 12 seconds, at 115 credits for 720p and 6 seconds. For a talking shot, keep the prompt specific: who speaks, the exact line in quotes, and the tone. "She says, quietly: 'I told you it would rain.'" works better than "she talks about the weather." Product shots and high resolution For product close-ups, the details matter: logos, edges, textures, labels. Resolution helps here. - Kling 3.0 from Kuaishou is a general-purpose model from text or image, with output from 720p up to 4K and optional audio. Clips are 5 or 10 seconds, and the default 720p, 5 second run is 155 credits. - FLUX 3 from Black Forest Labs makes text or image to video with sound, at 720p or 1080p, for 5, 10 or 20 seconds. The default 720p, 5 second run is 315 credits. Start product shots from a clean still of the product, ideally made with an image model first, then ask for one slow camera move: "Slow orbit around the bottle, studio light, label stays sharp and unchanged." Long takes Some ideas need one continuous shot: a walk through a room, a slow reveal, a scene that plays out without a cut. - Wan 3.0 from Alibaba runs up to 30 seconds in one take, with audio written in the same pass. Lengths are 5, 10, 15 or 30 seconds, at 480p to 1080p, with the 720p, 5 second default at 185 credits. - Seedance 2.5 makes long takes up to 30 seconds with generated audio, 480p to 1080p. The default 480p, 5 second run is 300 credits. - FLUX 3 also reaches 20 seconds if you want a long clip at 1080p. Long takes cost more, and they fail in more interesting ways, so draft the first few seconds short before committing to the full length. Drafting inside one family If you know the final will be on a Seedance model, you can draft within the family. Seedance 2 Fast is the quicker, cheaper Seedance tier for draft passes, at 480p or 720p and up to 10 seconds, with a 720p, 6 second default of 445 credits. Seedance 2 Mini is the lowest-cost route if sound in the draft is enough. A simple decision path 1. Not sure the idea works? Tanvo Video v1 or MiniMax H3. 2. Needs sound in a short clip? Seedance 2 Mini to draft, Veo 3.1 or Seedance 2.0 to finish. 3. Needs sharp detail? Kling 3.0 or FLUX 3. 4. Needs one long take? Wan 3.0 or Seedance 2.5. No map replaces a test. Run the same photo and prompt on two candidates at low resolution, compare them, and pay for full quality once. If a run fails, the credits are refunded. Browse every model on the video page. ### Seven prompt patterns for camera movement https://tanvo.ai/blog/prompt-patterns · Guides · Sep 18, 2026 Camera movement is where many AI video prompts go wrong. Writers describe the scene in detail, add "cinematic camera" at the end, and leave the model to guess. Models respond much better to plain, specific instructions in film language: what the camera does, how fast, and where it ends up. Here are seven patterns that cover most shots, with phrasing you can copy and adapt. Before you start: three rules - One camera move per shot. "Push in, then orbit, then crane up" in six seconds rarely works. If you need three moves, make three clips. - Put the camera in its own sentence. Describe the subject's action first, then the camera: "He pours the coffee. Slow push-in toward the cup." - Give a speed. "Slow," "steady" and "fast" change the result more than most adjectives do. 1. Locked-off The camera doesn't move at all. This is the most underrated pattern, because it gives all the motion budget to the subject and is the easiest way to keep a face stable. Phrasing: "Static shot, camera locked off on a tripod. No camera movement." Use it for: talking shots, reactions, dance clips, anything where the person matters more than the space. Say "no camera movement" explicitly, because some models add drift by default. 2. Slow push-in The camera moves steadily toward the subject. It adds tension or intimacy without changing the composition much. Phrasing: "Slow push-in from a medium shot to a close-up on her face." Naming the start and end framing ("from a medium shot to a close-up") gives the model a clear target. Without the end point, push-ins sometimes stop halfway or overshoot into an awkward crop. 3. Pull-back reveal The opposite: start close and move away to show where the subject is. Phrasing: "Start on a close-up of his hands on the steering wheel. Camera pulls back slowly to reveal the whole car on an empty road at dusk." Reveals work best when the start frame is tight. If you're working from a photo, the photo should be the close-up, and the prompt describes what appears as the camera pulls away. 4. Orbit The camera circles around the subject while staying pointed at it. It works well for products, outfits and hero moments. Phrasing: "Camera orbits slowly around the subject, a half circle from left to right, keeping her centered in frame." Say how far the orbit goes. "Half circle" or "quarter turn" is easier to follow than "orbits around," which can turn into a fast spin. For products, add "the label stays sharp and readable." 5. Tracking shot The camera moves alongside or behind a moving subject at the same speed. Phrasing: "Tracking shot. The camera follows behind her at shoulder height as she walks down the corridor, keeping the same distance." "Keeping the same distance" is the important part. Without it, the subject often walks away from the camera or the camera catches up and passes them. 6. Whip-pan A very fast horizontal pan that blurs the frame. It's often used as a transition between two moments. Phrasing: "He looks to the left. Fast whip-pan to the left with motion blur, landing on the open doorway." Always say where the pan lands. A whip-pan with no destination tends to blur into nothing and come back to the same shot. 7. Crane or tilt up The camera rises or tilts upward, often to end a scene or show scale. Phrasing: "Camera cranes up slowly from street level to a high angle, looking down on the crowd." For a tilt from a fixed point: "Camera tilts up from her shoes to her face." Tilts are easier than cranes for most models, because the camera position stays the same. If a crane shot keeps breaking, try the tilt version first. Putting a pattern into a full prompt A complete prompt usually has four parts, in this order: 1. Subject and action, as beats: "She stands at the window. She turns to face the camera and smiles." 2. Camera: "Slow push-in from a medium shot to a close-up." 3. Light and setting: "Soft morning light, plain white room." 4. What stays the same: "Her face, hair and clothes stay unchanged." Keep it short. If the camera line is buried in the middle of a long description, it's more likely to be ignored. Common mistakes - Conflicting instructions. "Static shot" plus "camera follows her" will give you one or the other at random. - Asking for a move the duration can't hold. A full 360-degree orbit in four seconds will look rushed. Ask for a half circle, or choose a longer clip. - Moving camera, moving subject, moving background. If everything moves, faces drift. Lock one of the three. Testing your phrasing Camera prompts are cheap to test at low resolution. Run the same prompt twice on a draft model such as Tanvo Video v1 or MiniMax H3, and if the move comes out the same both times, your phrasing is clear enough to take to a higher-quality model. If it comes out differently each time, the prompt is still ambiguous. Usually the fix is a speed, a start and end framing, or a destination. You can try all seven patterns on any model from the video page. ### A credit-saving workflow for longer projects https://tanvo.ai/blog/credit-saving · Guides · Sep 12, 2026 A single clip is easy to budget. A longer project, such as a short story told in eight shots, a product launch with a dozen cutaways or a music video, is where credits disappear. Usually it isn't the final renders that eat the budget. It's the experiments run at full quality that never make the edit. The fix is a simple order of work: plan on paper, solve the look in stills, solve the motion in cheap drafts, and only then spend on the final. Prices below are each model's default setting at the time of writing; the exact price for your settings is shown before every run, and the pricing page has the details. Step 1: Write the shot list first Before you generate anything, write every shot as one line: what we see, what moves, which camera move, and how long it lasts. - "Shot 3: close-up of hands opening the box. Locked off. 5 seconds." - "Shot 4: she lifts the watch and smiles. Slow push-in. 6 seconds." This takes ten minutes and saves more credits than any other step. It stops you from generating shots you don't need, and it tells you which shots need sound, which need a long take and which can be short. Step 2: Solve the look in stills Images cost a fraction of video. Every question about how the project looks, whether that's the character, the outfit, the light or the room, can be answered with stills before a single frame moves. At the time of writing: - Tanvo Image v1, the house engine, makes 1K images from text at 5 credits. - GPT Image 2.5 and Qwen Image 3 start at 10 credits at 1K. Qwen Image 3 is built for readable text and layout, so use it for title cards and anything with lettering. - Nano Banana 2 Lite and Seedream 5.0 Lite edit at a flat 15 credits. - Nano Banana 2 works from up to 14 photos at 25 credits for 1K, which helps when you need the same person across many frames. Make one start frame per shot. These become the first frames of your video runs, which means the look is already settled before you pay for motion. Step 3: Draft the motion at 480p Now animate each start frame at low resolution. The goal of a draft is to answer three questions: does the motion happen in the right order, does the camera do what you asked, and does the face hold? Good draft options, at their defaults: - Tanvo Video v1: 480p, 5 seconds, 20 credits. - MiniMax H3: 480p, 6 seconds, 90 credits. - Seedance 2 Mini: 480p, 6 seconds, 130 credits, with sound. Keep drafts short. If the final shot is 10 seconds, a 5 second draft usually tells you whether the first beats work. Fix the prompt, not the resolution: if a draft looks wrong, it will look wrong at 1080p too. Step 4: Lock the shot When a draft works, write down exactly what produced it: the start frame, the full prompt, the model, the duration and the aspect ratio. Don't touch the prompt again. Small rewording before the final run is the most common way to lose a working shot and pay twice. A simple way to do this is a text file with one block per shot: 1. Start frame: shot04_v3.png 2. Prompt: copied word for word 3. Model and length: as drafted 4. Status: locked Step 5: Spend once at full quality Only now do you move each locked shot to the model and resolution you want for the final. Some examples of where finals often go, at their defaults: - Veo 3.1 for short shots with sound: Lite at 720p, 6 seconds, 115 credits. - Kling 3.0 for detail, up to 4K: 720p, 5 seconds, 155 credits. - Wan 3.0 for long takes up to 30 seconds: 720p, 5 seconds, 185 credits. - Seedance 2.0 for native audio up to 1080p: 720p, 6 seconds, 480 credits. Higher resolutions and longer durations cost more than these defaults, so check the price shown before each run. More ways to save - Use cuts instead of long takes. Three 5 second shots are often cheaper and easier to control than one 15 second take, and an edit hides small problems. - Generate only the length you'll use. If the edit needs 4 seconds, don't render 10. - Don't upscale experiments. Only locked shots go to high resolution. - Re-roll the start frame, not the video. If a shot keeps failing, the problem is often the first frame. A new still costs far less than another video run. - Failed runs are refunded. If a generation errors out, the credits come back, so there's no need to avoid a model because you're worried about a failure. A sample budget shape For an eight-shot project, a sensible split is a handful of stills per shot, two or three short drafts per shot, and one final per shot. Most of the runs happen on the cheap engines, and the expensive models only see shots that already work. Start with the stills on the image page, then move to the video page. ### Keeping one character across ten shots https://tanvo.ai/blog/consistent-characters · Guides · Aug 27, 2026 One good AI clip of a character is easy. Ten clips where that character is clearly the same person, with the same face, the same hair and the same jacket, is much harder. Every generation is a fresh guess, and small differences add up across shots until shot ten looks like a cousin of shot one. You can't remove drift entirely, but you can make it small and predictable. This is the workflow that works best. Build a reference set first Before any video, make a small set of still images of your character. Think of it as a casting sheet: - A front-facing portrait, neutral expression, even light. - A three-quarter view. - A profile. - A full-body shot in the main outfit. - Optionally, the same face smiling and serious. If the character is based on a real photo, start from that. If they're invented, generate the first portrait, pick the one you like, and build the rest of the set by editing from it rather than starting over each time. Nano Banana 2 can work from up to 14 photos at once, which makes it a good fit for this: feed in the portraits you already have and ask for the next angle. Keep the set consistent itself. If the hair is different in the profile, fix it now, not after ten videos. Write a fixed description block Write one short paragraph that describes the character, and paste it into every prompt, word for word: "A woman in her thirties with short black hair, a small scar above her left eyebrow, silver hoop earrings, a mustard-yellow raincoat over a grey sweater." Rules for the block: - Be specific about what's distinctive. "Short black hair" is weak. A scar, an earring and a coat color give the model things to hold on to. - Never reword it. Changing "mustard-yellow" to "yellow" in shot six is an invitation to change the coat. - Keep it short. A long block competes with the action and the camera for the model's attention. Pin the first frame The strongest tool for consistency in video is image-to-video with a start frame you control. Instead of asking a video model to invent the character from text, give it a still of the character already in the shot. For each shot: 1. Make the start frame as a still, using your reference set, in the right setting and pose. 2. Check the face against the front portrait. If it's off, fix the still now. 3. Animate the still with a short prompt: the beats, one camera move, and a closing line such as "Her face, hair and raincoat stay unchanged throughout." This moves the hard part, getting the face right, into the cheap image step. The video model only has to keep what it's been given. Shot design that helps Some shots are simply easier to keep consistent than others: - Shorter shots drift less. A face that holds for 5 seconds may wander by 10. Cut more often. - Keep the face in view. A character who turns fully away and turns back is effectively regenerated. If they must turn, cut on the turn. - Limit camera moves near faces. A locked-off or slow push-in shot keeps a face steadier than an orbit. - Light the same way. A big change of light, from daylight to neon for example, changes how the model reads the face. Keep a lighting note in the description block if the shots belong to one scene. When to give up and re-roll Some runs just drift. The skill is knowing when to stop fixing and start over. - If the face changes in the first second, the problem is the start frame or the prompt. Re-roll the still, not the video. - If it changes late in the clip, shorten the clip or cut before it happens. - If two re-rolls of the same shot both drift, change something: a tighter crop on the face, a simpler action or a different camera move. A third identical attempt rarely does better. - If the outfit changes but the face holds, add the outfit to the closing "stays unchanged" line and make sure the outfit is clearly visible in the start frame. Draft on a cheap model first. At the time of writing, Tanvo Video v1 makes 480p, 5 second drafts at 20 credits, and Seedance 2 Mini starts at 130 credits for 480p, 6 seconds. Once a shot holds, take the same start frame and prompt to your final model. Failed runs are refunded, so a run that errors out costs nothing. A checklist for each shot - Start frame made from the reference set and checked against the portrait. - Description block pasted in, unchanged. - Two or three beats, one camera move. - A closing line naming face, hair and clothes. - Short duration, with a cut planned where the character turns away. Start with your reference set on the image page.