Documentation
Read the docs.
Detailed documentation lives in the GitHub repository, so it stays in sync with the code. Below are jump-off points to the most useful pages.
Self-hosting guide
docker compose up. Postgres + Redis + MinIO + Ollama on your hardware. No paid API required.
Configuration reference
Every environment variable, every quota, every model endpoint — what it does and what to set it to.
Template authoring guide
How to build a reusable extraction schema with the visual field-picker. Anchors, types, versioning.
API reference
REST endpoints, authentication, rate limits, webhooks, error codes. Auto-generated OpenAPI spec.
Extraction pipeline
Deep dive on the OCR cascade, extraction methods, grounding rules, confidence math, and human-review boundary.
Supported languages
Per-language maturity tier, OCR engine path, accuracy notes, known caveats.
Architecture deep-dive
How the web app, engine, workers, queue, and inference layer fit together.
Known issues
Honest list of current limitations. Tracked openly so you know what to validate before committing.