Get running
Start with Docker Compose, then finish the first-run model configuration in the web UI.
Get running
Start with Docker Compose, then finish the first-run model configuration in the web UI.
Configure models
Add provider connections and assign separate models for metadata extraction, grounded Q&A, and OCR.
Troubleshoot OCR
Learn when to use LLM OCR, when to keep OCR local with Tesseract, and how to tune the OCR path.
Understand the repo
See the development commands, repository layout, and where the backend, workers, and UI live.