Tanvo
5

Developers

Build with Tanvo

Images, video with sound and full songs from your code or your AI assistant: 30 models, 80 ready-made apps, one API.

OverviewAPI referenceMCP serveropenapi.yaml

Quickstart

One request makes an image. Without an API key it runs on the free tier (the house image engine, watermarked), so you can try every snippet below right now.

# Free tier: make up any id for x-anon-id (no account needed)
curl -s https://tanvo.ai/api/v1/generations \
  -H "content-type: application/json" -H "x-anon-id: demo-$(uuidgen)" \
  -d '{"kind":"image","model":"studio-image-v1","prompt":"a paper lantern over a misty lake at dawn",
       "options":{"aspect":"16:9","resolution":"1K","format":"PNG"}}'

# Then poll until status is SUCCEEDED (same x-anon-id)
curl -s https://tanvo.ai/api/v1/generations/<id> -H "x-anon-id: demo-…"

Ways to build

REST API

Submit a run, poll it or get a signed webhook, download the file. JSON over HTTPS with CORS.

API reference →

MCP server

Every app and model inside Claude, Cursor, Codex, Gemini CLI and other MCP clients.

Set up MCP →

Python SDK

pip install tanvo. No dependencies, typed, retries that never double-charge.

tanvo-python on GitHub →

JavaScript / TypeScript SDK

npm install @tanvoai/sdk. Node, Deno, Bun, edge runtimes and browsers.

tanvo-js on GitHub →

Keys and the free tier

IdentityHowWhat it can run
API keyHeader Authorization: Bearer sk_…. Create one under Settings → API keys; it is shown once.Every model, your plan's limits and prices, clean output; runs land in your history.
Free tierHeader x-anon-id with any id you make up (8–80 characters of letters, digits, . _ -).The house image engine only, watermarked: 5 credits per id (one image) and a few ids per address per day.

Keep keys on your server. A key carries your account's credits, so never ship it in a browser bundle or a mobile app; call the API from your backend instead.

Ready-made apps

Beyond raw models, Tanvo has 80 apps (pet portraits, figurines, old photo restoration, product shots, trend videos, birthday songs), each with preset looks: a tuned prompt, model and settings with a real example output. GET /api/v1/apps lists them, and both SDKs and the MCP server run a look on your photos in one call.

portrait = client.generate_from_app(
    "ai-pet-portrait-generator", look="royal", photos=["rex.jpg"], details="Rex, a very good boy"
)

Credits and pricing

API runs use the same credits and prices as the website: 14 image models, 15 video models and Suno V6 for songs. Every model's price at its default settings is in GET /api/v1/models (credits_at_defaults), and the exact cost of a run is in its cost field. Failed runs are refunded automatically, and a repeat with the same requestKey is never charged twice. Plans and credit packs are on the pricing page.

Next steps

  • API reference: endpoints, the generation object, songs, webhooks, errors and limits.
  • MCP server: install in your assistant and the ten tools it adds.
  • openapi.yaml: the machine-readable spec, for code generators and API clients.
  • llms-full.txt: every model, app and article in one plain-text file for LLMs.
  • Questions: support@tanvo.ai