Most document workflows are several steps — convert, extract, clean up, check. A pipeline chains those processors into one versioned object you call by name, instead of orchestration you wire and debug by hand.
Drag together the processors a document needs — convert, extract, your own code — and chain them in the playground until the output is right. Each one is backed by our open models: Marker, Surya, and Chandra. Save your work and you get a versioned pipeline you can keep iterating on.
Publish a pipeline version and you get an immutable ID. Run it from the SDK, the REST API, a notebook, or CI — same pinned version, same output everywhere. Move between dev, stage, and prod with named environments.
Create a rubric and run evals on pipeline outputs periodically — quality improves as we ship model updates, and any regression gets flagged before it reaches production.
The processors run on Marker, Surya, and Chandra — the same open models we publish and that teams already run in production. Read the source, benchmark them yourself, then build on top.
See the open-source models →A published version is frozen and addressed by ID, so a run today and a run next year resolve to the exact same steps. Reindex months later and the output still matches.
Build one in the playground →Start with an API key — nothing to host or operate.
Sign up →Run in-region, with no egress to US infrastructure.
Sign up →Runs inside your own AWS, GCP, or Azure account.
Talk to sales →Fully offline, the same model weights, dedicated support.
Talk to sales →Free tier with up to $20 in credits per month — no card required.
