Open-Source PDF Table Extraction with Camelot: Quick‑Start Guide

In today’s information‑overload environment, professionals often need to pull tabular data out of PDFs such as financial reports, market research, or legal documents, but manual extraction is slow and error‑prone.

The article introduces Camelot, an open‑source Python library designed to parse PDF tables and export them as CSV files.

Camelot quick start

Create a new Camelot project directory.

Install Ghostscript, which Camelot relies on to read PDFs. macOS users can run brew install ghostscript.

Install Camelot via pip: pip install "camelot-py[base]" Create a main.py file with the following code:

import camelot

tables = camelot.read_pdf('foo.pdf')
tables.export('foo.csv', f='csv', compress=False)

Run the script with python3 main.py. If the program aborts with a RuntimeError about Ghostscript, set the DYLD_LIBRARY_PATH environment variable, for example:

export DYLD_LIBRARY_PATH=/opt/homebrew/Cellar/ghostscript/10.03.1/lib/

After fixing the path, re‑run the script; a CSV file will appear in the project root.

Excalibur quick start

To simplify usage, the Camelot team provides Excalibur, a web‑based UI.

Install Excalibur with pip install excalibur-py.

Initialize its SQLite database: excalibur initdb Start the server: excalibur webserver Then open http://127.0.0.1:5000/files in a browser.

In the browser UI, click the “Upload PDF” button, select a local file, and Excalibur extracts the tables, also offering a visual table‑detection preview.

https://github.com/camelot-dev/camelot