In development. Built for real paper.
Give your
documents
a reader.
The invoice in your inbox. The handwritten register. The PDF that’s really a photo. Turn them into data you can work with.
Readable data. Visible uncertainty. Less retyping.
Paper in. Possibilities out.
businesses actually use.
Invoices & vouchers
Handwritten records
Images & scanned PDFs
Different documents.
Different challenges.
A sharp scan and a handwritten logbook need different readers. KenOCR brings specialist engines together behind a common extraction API.
Document illustration
Keep the structure.
Keep the detail.
Extract what’s written into rows and cells. Keep empty fields distinct from dashes, and bring uncertain readings to the surface.
Dedicated document-specific APIs are planned. The current preview uses one extraction endpoint.
A second look beats
a confident guess.
When two independent readings disagree, you should know. KenOCR returns flags alongside the extracted data, so a plausible answer doesn’t quietly become your next data problem.
Rows What was read
Flags What needs attention
Proof What was cross-checked
A document in.
Data back.
Send a file to the extraction endpoint. Get structured output and an explicit account of uncertainty. Use the data in your own app, workflow or business system.
Read the API guideDeveloper preview. A KenOCR server and API key are required. Public hosted access and pricing are not available yet.
curl http://localhost:8710/v1/extract \
-H "X-API-Key: $KEN_API_KEY" \
-F "file=@document.pdf" \
-F "mode=balanced"Where we’re headed
Expert readers.
Everyday economics.
We’re building towards fast, affordable OCR with dedicated expertise for invoices, handwriting and image-based PDFs. One familiar way to work with all of them.
