Turn a prompt into a hosted clip, billed by the second
Text-to-video and image-to-video across every major video model through one endpoint. Billed per second of output in USDC, quoted before generation starts.
AI video generator API: text to video and image to video from one endpoint
Send a prompt — and optionally a reference image, or a first and last frame — with a model, a duration and a resolution; the response is a job you poll. Some models are metered by the second at a flat rate, others on the tokens they generate, which scales with resolution; the card quotes each model's default clip so they compare.
The finished clip is mirrored to durable storage and served from a URL that does not expire, so the poll result can be handed straight to whatever consumes it.
How to call the video generation API — prompt, reference image, duration, resolution
One endpoint, priced in dollars before it runs and settled in USDC. No API key and no account — the request pays for itself.
POST /api/v1/videos/generations
Host: blockrun.aiAI video generation API pricing, per second and per clip, per model
AI video generation for agents: from an MCP host, a pipeline or the SDK
In Claude Code, Claude Desktop, Cursor, Windsurf, Cline and any other MCP host, this is the blockrun_video tool of the BlockRun MCP server — the agent calls it like any other tool and the result comes back in the conversation. In OpenClaw, Codex and Hermes, ClawRouter exposes the same catalog. In your own code, the Python and TypeScript SDKs wrap the endpoint above with types; any HTTP client works too. Every harness, with its setup →
No account and no API key stand between an agent and the first call: the request is quoted in dollars and the agent's client pays it. A team that prefers keys gets one, with a monthly invoice.
Video generation API vs Runway, Luma, fal.ai and Replicate
Runway and Luma sell video generation as a product with an API attached: a subscription with credits, their own models, their own player. fal.ai and Replicate host third-party video models behind a key and a prepaid balance. This page is a third shape: every major video model — Seedance, Sora, Grok Imagine and the rest of the card — behind one request, each priced per second of output before the job starts.
The models meter differently and the card says so. Some bill a flat rate per second at a given resolution; some bill on the tokens they generate, which rises with resolution and length. The card quotes each model's default clip so they compare on one line, and the 402 quotes the exact figure for the clip you actually request.
Generation is asynchronous everywhere: the call returns a job you poll, and the finished clip lands on a hosted URL that does not expire, so a pipeline can hand it straight to the next step — a voice track, a thumbnail, a publish — without storing the bytes itself.
AI video generator API questions: models, clip length, image-to-video, pricing
Which AI video generator models can I call through the API?
Every model on the rate card above — Seedance, Sora, Grok Imagine and more — through one video generation endpoint, each row linking to the model's page with its resolutions and duration ceiling.
Does the video generation API do image-to-video as well as text-to-video?
Yes. Every model on the card accepts a reference image as well as a prompt, and several take a first and last frame to animate between.
How is AI video generation priced?
Per second of generated video at the resolution you request, or on the tokens the model generates where that is how the vendor meters — the card quotes each model's default clip, and the 402 quotes your exact clip before the job starts.
How long can a generated video be?
It depends on the model. Each row on the rate card links to the model's own page with its duration ceiling and the resolutions it renders.
Is there a free AI video generation API?
No — every video model on the card is paid per clip. The free tier covers chat models. A video costs what its row says, quoted before generation begins.
How do I get the finished video from the API?
Generation is asynchronous: the call returns a job you poll, and the finished clip is served from a hosted URL that does not expire. Nothing has to be downloaded and re-hosted.
Can the video generation API produce 4K or vertical video?
Resolution and aspect ratio are per model — the model's page lists what it renders. Higher resolutions cost more on the token-metered models, and the quote reflects it.
Video generation API vs Runway, Luma, fal.ai or Replicate?
Runway and Luma sell their own models on a subscription; fal.ai and Replicate host others' behind a key and a balance. This page puts every major video model behind one request, priced per second before the call, with no account and no credits to pre-buy.
Can an AI agent generate video without an account?
Yes. The request pays for itself, so an agent with a funded balance can start a generation, poll it and collect the clip with no human signup.