Hey everyone! We built ReCognition because we kept hitting the same wall: every OCR API we tried forced us into their predefined fields. Need to extract something they didn't think of? Too bad, train a custom model or parse raw text yourself. So we built an API where you define the schema. Any document, any fields, structured JSON back. Need patient_name and test_results from a lab report? Just put that in your schema. Need container_weight from a customs form? Same thing. We also ship pre-tuned schemas for invoices, receipts, and purchase orders if you just want something that works out of the box. On the security side — we don't store your files. Everything is processed in memory on EU servers (Frankfurt) and wiped after delivery, and the location can be adjusted. For teams that can't send data to the cloud at all, we deploy on your infrastructure. It's in open beta — free to use, and beta users lock in 50% off for life when we go paid. Would love your feedback. What document types would you throw at it first?

ReCognition

Define extraction schema, send document, get structured JSON

Define extraction schema, send document, get structured JSON