doc2json
Turn PDFs and photos into typed JSON with your own schema. Missing values come back null.
Hosted MCP Server
npx add-mcp 'https://doc2json.manoonlabs.com/mcp'Installs into Claude Code, Codex, Cursor and more
Documentation
What kinds of documents can I send?
PDF, PNG, and JPEG. Free allows 10 MB and 10 pages per file; paid plans allow 25 MB and 50 pages. Invoices, receipts, statements, forms, résumés, scans, and phone photos all work. If a person can read it, doc2json can structure it.
Do I have to write a schema?
No. Without a schema, the model returns the fields it finds. A schema makes the output predictable: you get exactly your keys, in your types, every time.
What happens when a value is not in the document?
It comes back as null. doc2json is built to return null instead of guessing, so you can tell missing data from real data.
Do you store my documents?
No. The file is read for the parse and then dropped. The activity log keeps metadata only: status, duration, the API key, and the schema name. Never the document or the extracted text.
Can I get the results as Excel or CSV?
Yes. In the Playground, use Download after a parse. Over the API, send export=xlsx or export=csv with the parse and the response adds the file next to the JSON, for the same credits.
How do I use it with AI agents?
doc2json is an MCP server. Add it to your agent, and it gets a parse_document tool that uses your saved schemas.
Can I try it without an account?
Yes. The Playground, the schema editor, and the API docs are open to everyone. Sign in when you want an API key or want to parse your own files.