Live request & response schemas for every AI model API▌
Which models exist right now, and what their payloads look like. JSON Schemas for 7 providers, refreshed automatically, served as plain JSON over HTTP — no key needed for reads.
try it
$ curl https://modelschemas.com/v1/models?q=claude
{ "id": "anthropic-claude-sonnet-5", "rawId": "claude-sonnet-5", "lastSeenAt": 1785082565, … }live7 providers2,064 models1,394 endpoints2,770 schema versionspolled 11m ago · specs synced 11h ago
Providers
GET /v1/providers →| provider | status | models | endpoints | schemas | polled | synced |
|---|---|---|---|---|---|---|
| Anthropic | active | 13 | 5 | 10 | 11m ago | 11h ago |
| ElevenLabs | active | 10 | 34 | 58 | 11m ago | 11h ago |
| FAL | active | 1,434 | 1,262 | 2,524 | 11m ago | 2d ago |
| Google Gemini | active | 59 | 19 | 38 | 11m ago | 11h ago |
| xAI Grok | active | 11 | 12 | 24 | 11m ago | 11h ago |
| OpenAI | active | 129 | 38 | 69 | 11m ago | 11h ago |
| OpenRouter | active | 408 | 24 | 47 | 11m ago | 11h ago |
Recent changes
GET /v1/changes →| 11m ago | model.updated | Model openai/gpt-5.1-codex updated | openrouter |
| 2h ago | model.updated | Model z-ai/glm-5.2 updated | openrouter |
| 3h ago | model.updated | Model z-ai/glm-5.2 updated | openrouter |
| 3h ago | model.updated | Model tencent/hy3-preview updated | openrouter |
| 3h ago | model.updated | Model z-ai/glm-5.2 updated | openrouter |
| 4h ago | model.updated | Model z-ai/glm-5.2 updated | openrouter |
| 4h ago | model.updated | Model z-ai/glm-5.2 updated | openrouter |
| 5h ago | model.updated | Model z-ai/glm-5.2 updated | openrouter |
Subscribe with a webhook to get these pushed — POST /v1/subscriptions
Examples
/examples →| schema-studio | generative UI — request forms rendered at runtime from live provider schemas | @modelschemas/vite |
| image-dimensions | every image model with its supported output sizes, extracted from each input schema | @modelschemas/client |
| video-composer | video request builder whose knobs only offer values the schema allows | @modelschemas/client |