Get document loading right.
Route every page, trace every step and check what reaches your model. complydoc reads your documents the way your pipeline does, and shows you what it got wrong.
- Runs on your machine
- No account, no key
- MIT licensed

How the core works
Three stages run on every document, on your machine. Each one writes what it decided and why into the report, so a run can be traced and repeated.
Each page goes the cheapest way that reads it
The router measures every page before anything is sent. A page with a good text layer is read as text, a scan goes to local OCR, and only a page with charts or handwriting goes to a vision model. Hover a page to see why it went where it did.
- annual-report-2025.pdf
- employee-handbook.pdf
- master-services-agreement.pdf
- rechnungen-2026-de.pdf
- supplier-invoices-scanned.pdf
- vendor-due-diligence.pdf
routed $0.048 · every page to Claude Sonnet 5 as an image $0.106
Personal data and hidden instructions, found before indexing
Deterministic rules find identifiers such as tax numbers, accounts and emails, and verify their checksums. Text a person cannot see, like white or tiny type, is taken out, and instructions written to an AI are flagged. The network is blocked while this runs.

Every step written down, page by page
The report keeps the text the loader returned for every page, with each finding, route and timing. Open it in the viewer, diff two runs, or fail a build in CI when ingestion gets worse.

Measured, and rerunnable
Detection is scored on a labelled corpus that ships with the package. The router's rules each won on a public benchmark.
- 53 of 53
- labelled identifiers found, across 39 categories
- 0
- wrong flags, on passages full of invoice and order numbers
- 51
- passages in the corpus, shipped with the package
See what your model would be given.
complydoc's real output on two sample documents. Nothing here is mocked up.
Same page, two loaders, two different texts.
PyPDF reads each column in turn. PDFPlumber reads straight across both and splices the clauses together, with no error. Compare yours
master-services-agreement.pdf, page 2
- PyPDFLoader kept clause 1.1 whole
- PDFPlumberLoader split clause 1.1 with the other column
Flag malicious attempts on your documents.
This CV hides an instruction to the model in a line set at half a point. complydoc finds it and says why it was flagged. How hidden text is found
cv-data-engineer.pdf, as a person sees it and as the model reads it
Works with your existing loaders
complydoc reads documents by their shape, so it needs no adapter per loader and imports no framework you do not use. Pass it a loader, a folder or a list of documents.
Any loader with load(), load_data() or lazy_load() works, as does any callable that returns documents.
Run it on your own folder
complydoc runs locally. Network access is blocked while documents are read, and each report records that. MIT licensed.
complydoc ui opens every report and pipeline run in the folder in your browser, served from your machine.




