Describe the table you want
Create a schema with field names, types, and examples—plus list fields for repeating records. It becomes the blueprint for every document in the project.
Define the information you need—down to every line item—upload your PDFs, and get a clean, editable table in minutes, not hours of copy and paste. Or hand the job to your AI assistant or your code.
get_project_dataCreate a schema with field names, types, and examples—plus list fields for repeating records. It becomes the blueprint for every document in the project.
Drop in files, paste public links, or send them from your assistant or code. PDF2Table reads each page, extracts the fields, and keeps the source close for review.
Edit any value right in the table, open its source PDF when you need context, then export everything to Excel.
NEW NESTED LIST FIELDS
Real documents repeat: charges on a card statement, products on an invoice, dependents on an insurance policy. List fields capture each one as its own typed record—in document order, linked to the page it came from—so nothing gets flattened or left behind.
NEW MCP SERVER & API
PDF2Table speaks the Model Context Protocol. Paste one URL into Claude, ChatGPT, Cursor, or Claude Code, sign in, and your assistant can design schemas, process PDFs, and read the results—with the same projects, permissions, and credits as your team. Prefer code? Use an API key and call the same tools from any script or pipeline.
Pull every charge from these 12 card statements into our Expenses project.
Done — 12 statements, 412 charges. Every total reconciles except amex-march.pdf, which is $18.40 short, so I flagged it for review.
$ curl https://…/api/mcp \-H "Authorization: Bearer p2t_•••" \-d '{"method":"tools/call", …}'{ "documents": [{"file_name": "amex-march.pdf","data": { "charges": [{ "merchant": "AWS", "amount": 188.95 }, …
BUILT FOR REPEATABLE WORK
Keep invoices, statements, forms, and reports organized by project. Each project uses one schema, so every new file lands in a familiar table.
Vendor invoices from January–March
READY WHEN YOUR PDF IS
Create a workspace, define your first schema, and turn your next batch of PDFs into a table you can actually use—in the app, in your AI assistant, or straight into your code.
Start extracting