Boring API · available

PDF Tables

Extracts tables from text-based PDFs with page, bounding box and a quality level per table; detects header rows, titles and footnotes, merges tables continued across pages only with evidence; returns JSON or the best table as CSV.

Call

curl -F file=@price-list.pdf https://api.relaex-engineers.de/v1/documents/tables

Input: Multipart file (PDF up to 20 MiB and 60 pages), optional options (JSON): pages, strategy lines|text|auto, include_cells, output json|csv, table_id. Works for any language and script with embedded text.
Limits: 90 s per document; scanned pages without text are reported as OCR_REQUIRED.

For AI agents

MCP tool documents_tables over Streamable HTTP at https://api.relaex-engineers.de/mcp (no authentication). In Claude Code: claude mcp add --transport http boring https://api.relaex-engineers.de/mcp. In Claude.ai, ChatGPT and other clients add the same URL as a custom connector. OpenAPI-based frameworks use /openapi.json.

Languages

Responses are in English; error, warning and rule codes are language-neutral so your assistant can explain them in your language. Column headers and common cell words are recognised in English, German, French, Spanish, Italian, Portuguese, Dutch, Polish, Czech, Turkish, Swedish, Danish, Norwegian, Finnish, Korean, Japanese and Chinese; numbers in both 1.234,56 and 1,234.56 style; any ISO 4217 currency (no conversion).

Access and privacy

Open with a rate limit per IP and session. An X-API-Key raises the limit and unlocks GET /v1/usage/me. Files are processed in memory and never stored.

llms.txt · Health · API docs · Connect · Privacy