Skip to main content
31 tools. These are not models. A tool does exactly one thing, deterministically, for a fixed price: trim a video, extract text from a PDF, pack an archive, run a Python snippet. No prompt, no sampling, no variance between two identical calls. GET /v1/tools lists them at runtime — see the API reference. They exist because most real work around a model is not inference. A pipeline that captions a video needs the audio extracted before a speech model ever sees it, and paying inference prices for ffmpeg would be absurd.

​
Calling one

Every tool takes the same envelope — an id, and the input and params that tool declares:
Each tool page lists that tool’s parameters and gives runnable examples.