MakeAIVideo API documentation
Turn a brief or your own script into a finished short-form video, edit and regenerate parts of it, and fetch the MP4. Everything the product can do, agents can do.
https://app.makeaivideo.ai/api/v1 · v1.2 · 51 paths, 65 operations
Start here
Ways to connect
Three surfaces onto the same account, the same credits and the same scopes.
MCP server
Add MakeAIVideo to Claude or ChatGPT as a custom connector.
Agents
Async renders, idempotency, and least-privilege keys for autonomous agents.
CLI
Device-code login for terminals and headless machines.
TypeScript SDK
A typed client with polling, webhook verification and error classes.
Open source packages (MIT): @makeaivideo/mcp (GitHub, npm) · @makeaivideo/sdk (GitHub, npm) · @makeaivideo/cli (GitHub, npm). Machine-readable spec: openapi.json.
The response envelope
Every JSON response has the same three keys, so a client can unwrap any endpoint the same way.
{
"data": { ... },
"error": null,
"meta": { ... }
}On failure data is null and error carries a machine-readable code.
{
"data": null,
"error": {
"message": "Insufficient credits",
"code": "insufficient_credits",
"retryable": false
},
"meta": { ... }
}retryable is true only for 5xx and 429. Never retry a 4xx without changing the request. Two endpoints return files rather than JSON, including GET /videos/{id}/captions.
The core loop
Renders take minutes, so creating a video is asynchronous: you get a video_id immediately and collect the result later.
- 1
GET /toolsSee what we can make. - 2
POST /videos/estimateFree quote, before you spend credits. - 3
POST /videosCreate from a brief plus fields, or your own script. Returns 202. - 4
GET /videos/{id}Poll, respecting poll_after_seconds, or skip polling and subscribe to video.ready. - 5
GET /videos/{id}/downloadFetch the finished MP4.
Set idempotency_key on create so a retry never produces, and never charges for, a second video.
API reference
Generated from the OpenAPI document, so it matches what the API actually serves.
- All endpointsEvery endpoint in one list, grouped by resource.
- Introspection3 operations.
- VideosCreate, inspect, edit, regenerate, export and download videos.
- ScriptsWrite or structure a script before rendering.
- Characters7 operations.
- Voices3 operations.
- Music2 operations.
- Uploads4 operations.
- Brand kit3 operations.
- Workspace2 operations.
- Ideas2 operations.
- PublishingConnect social accounts and post finished videos to them.
- Webhooks7 operations.
- AuthCLI device-code login (RFC 8628). Unauthenticated.
API questions
- Does MakeAIVideo have an API?
- Yes. There is a REST API at https://app.makeaivideo.ai/api/v1 covering 56 operations, plus an MCP server for assistants. Anything the product does, the API does.
- Can I generate video programmatically?
- Yes. POST a brief or your own script to /videos and the API renders a finished video with narration, scenes, captions and music, which you then download as an MP4.
- How much does the API cost?
- The API is not billed separately: renders use the same credits and the same balance as the app. Every plan from $9 a month includes monthly credits, reset on each billing date, and every plan starts with a 7-day free trial. A 30-second stock-footage video is about 112 credits and a 30-second AI video with voiceover about 612. POST /videos/estimate quotes any job for free before you commit credits to it.
- How long does a render take?
- Minutes rather than seconds. Creation is asynchronous: you get a video id straight away and either poll for status or subscribe to a webhook that fires when the render is ready.
Get an API key
The API and MCP server use the same credits and the same balance as the app, with no separate API plan. Create a key under Settings, Advanced, API keys.
7-day free trial with 350 credits, then plans from $9/mo. Cancel anytime before the trial ends.