About

We're building the relational layer for the document web.

Most company data lives in files: spreadsheets, PDFs, exported reports, Word docs, scanned forms. The information is right there — but it isn't queryable. So teams build glue code, copy-paste, and burn weeks on plumbing that doesn't deliver value.

DocQuery flips that. Upload any document, and within minutes you have a real relational model exposed as REST, GraphQL, and MCP APIs. Your apps and your AI agents query the same source of truth.

What we believe

Documents are data. The only thing standing between a PDF and a SQL table is a parser. The state of the art is now good enough that anyone should have one.

Relational beats flat. Spreadsheets are easy to author but terrible to query at scale. The minute you have two related files, you need foreign keys — even if you've never heard the term.

AI agents need real APIs. Feeding a 50-page PDF into a chat context is wasteful and unreliable. Feeding a typed, queryable MCP endpoint is fast and verifiable.

Humans stay in the loop. AI extraction is amazing, and also occasionally wrong. We always show you the extracted data before it goes live, and we make corrections trivial.

Boring tech, ambitious surface. Phoenix, SQLite, Req. A handful of dependencies. The product is opinionated; the stack is dull on purpose.

Get in touch

Questions, feedback, or a complicated document we should be able to handle? We want to hear about it.