Who it is forAn analyst who wants the data tables from a report and not the prose.
The problemRetyping tables from a PDF is slow and error prone.
The outcomeTable content converted to markdown for the pages you selected.
What you need first
- Python 3.10+ and PyTorch
- pip
- A GPU is best, CPU and Apple Silicon also work in fast mode
Steps
- Install with
pip install marker-pdf - Run
marker_single /path/to/file.pdfon your report - Add
--converter_clswith Marker.converters.table.TableConverter to extract tables only - Add
--page_rangeto limit which pages are processed - Review the output and paste tables into your spreadsheet
Prompt for your agent
With datalab-to/marker, run `marker_single /path/to/file.pdf --converter_cls marker.converters.table.TableConverter --page_range "0,5-10,20"` on <my report PDF> and turn each extracted table into a CSV.
Caveat. Check numbers against the source; table reconstruction can fall back to a model and may err.
Use this in your AI agent
Ask your agent
Using the Deduped MCP, get the setup steps for datalab-to/marker and "Pull just the tables out of a PDF"
Add Deduped to Claude Code once
claude mcp add --transport http deduped https://dedupedai.com/mcp
Every Monday: the 5 hottest repos and exactly how to use them. Free, one email a week, unsubscribe anytime.
Get them freeShare: X LinkedIn Hacker News Email Copy link
Third-party repos we don't own. AI-written use cases can be wrong. Disclaimer