Project ◆ Hidden gem
gmft
conjuncts/gmft · Python · MIT
Pull tables out of PDFs into DataFrame with almost no setup.
gmft is a deliberately small PDF table extractor — no heavy ML stack, just a couple of optional backends — aimed at getting clean tables into pandas fast.
35.6 / 100 · Score
★ 542 Stars
+0 30-day growth
3 min Time to first run
Not required API key
Score
Trend 18
Activity 32
Quality 90
Community 22
Relevance 43
Setup at a glance
DeployPip
GPUNo GPU needed
API keyNot required
UI languagesen
Doc languagesen
Install
pip install gmft Deep dive
gmft extracts tables from PDFs straight into pandas DataFrames — the unglamorous step that eats days of every document-processing project.
Beyond DataFrames it returns the pieces you need for downstream work: markdown, LaTeX, HTML, CSV and JSON by extension, plus the raw text with positions and a cropped image of each detected table.
It ships notebooks for both single-file quick starts and bulk extraction, and full documentation on Read the Docs.
Use cases
Scrape tables from research papers and reports
Automate data entry from PDF invoices
Quick EDA on tabular PDF dumps
Good for / Not ideal for
Good for
Data scientists who hate PDFs
Notebook-first workflows
Projects that must stay lean
Not ideal for
Scanned image-only PDFs (needs OCR)
You need full document structure, not just tables