DocQuery
Turn spreadsheets (Excel, CSV, Google Sheets, Airtable) and PDFs into one relational REST/GraphQL API and a remote MCP server. Query rows across files with filters and joins. Try it read-only with the demo key dq_demo_readonly.
Serveur MCP hébergé
npx add-mcp 'https://docquery.io/api/v1/mcp'S’installe dans Claude Code, Codex, Cursor et plus
Documentation
REST · GraphQL · MCP for AI agents
Your spreadsheet and the PDFs beside it. One relational API.
Connect the Excel workbook or Airtable base your team already lives in, add the invoices, contracts, or reports that go with it, and DocQuery links them into real tables with foreign keys. Query everything over REST, GraphQL, or MCP. When the sheet changes, the API follows.
Start free — no card needed See an example →
Free tier includes 5 documents and 10,000 rows. No credit card required.
1 · The living sheet
customers (Excel or Airtable)
id name plan
1 Acme pro
2 Globex starter
Stays connected: edits in the sheet flow into the table, and API writes can flow back.
2 · The PDFs beside it
invoices (extracted by AI)
number customer_id total status
1001 1 $4,200 paid
1002 2 $1,150 due
Columns, types, and keys are detected for you. Link customer_id → customers.id in one click.
3 · Ask across both
From Claude, Cursor, or your app
"Which pro customers have unpaid invoices?"
Your agent calls the MCP tools query_table and related_rows over the linked tables. Apps get the same data with ?include=customer_id on REST.
Try it now · no signup
Call the live API with the public demo key
A read-only workspace with the example customers and invoices above, already linked.
curl "https://docquery.io/api/v1/tables/example_invoices/rows?include=customer_id" \
-H "Authorization: Bearer dq_demo_readonly"
Same key works for MCP at /api/v1/mcp (setup). More calls are on the API quickstart. Your own workspace gets a private key with full read/write.
Not just a sheet-to-API.
Plain spreadsheet-to-JSON tools give you one flat table. Inbox parsers give you fields from one PDF at a time. The useful answers are in between: which contract goes with which customer, which invoice belongs to which project.
- Spreadsheets and documents in one workspace, with real foreign keys between them.
- Live sync with Excel (Microsoft 365), Google Sheets imports, and Airtable.
- The same data over REST (OpenAPI), GraphQL, and an MCP server for agents.
- Export any table to CSV or JSON. Your data never gets stuck here either.
- 1
Upload or connect
Drop in a CSV, JSON, Markdown, PDF, DOCX, or PPTX — or connect Airtable, Google Sheets, or Microsoft 365 with OAuth. - 2
Extract
AI detects tables, columns, types, and primary keys. You review and tweak. - 3
Relate
Link columns across tables with foreign keys — no SQL, no migrations. - 4
Ship
REST, GraphQL, and MCP endpoints are live the moment you save.
For apps + integrations
CRUD endpoints per table with filter, pagination, and relation expansion via ?include=. Auto-generated OpenAPI spec.
GET /api/v1/tables/invoices/rows?status=paid&include=customer_id
For flexible queries
Workspace-aware schema with nested relations. Filter, paginate, and traverse foreign keys in a single request.
{ rows(table: "invoices") {
data
related(column: "customer_id") {
data
}
} }
For AI agents
Native Model Context Protocol — Claude, Cursor, and any MCP-aware client gets tools, resources, and prompts out of the box.
POST /api/v1/mcp
{"method":"tools/call",
"params":{"name":"query_table",...}}
Analysts & ops teams
Stop copy-pasting between Excel and dashboards. Upload reports, link them, and query them over REST, GraphQL, or MCP.
Finance & accounting
Invoices, receipts, and bank statements turn into a real ERP-style API in minutes.
Legal & compliance
Contracts and supporting docs become related tables you can query.
AI engineers
Feed clean relational data to agents via MCP. No more pipeline-of-pipelines to get usable context.
SMB founders
Mix Google Sheets and supplier PDFs into one source of truth that your tools can read.
Enterprises
Pro raises the document and row caps. Team workspaces, audit logs, and row-level security are coming soon.
Stop building pipelines. Start querying.
Set up your first document → API in under five minutes. Free forever for personal projects.