Self-learning document intelligence
Any document in.
Canonical JSON out.
Upload a document and get a reviewable extraction with per-field confidence and provenance. Every human correction teaches the template — the more you review, the less you have to.
- Any layout — orders, invoices, IDs
- OCR for scanned pages
- Per-field confidence & provenance
- Self-learning from every review
- Headless /v1 API & embeddable review