Skip to content

FAQ

Jump to bottom
Nick edited this page Dec 18, 2025 · 1 revision

Frequently Asked Questions

Common questions about Chart2CSV.

What is Chart2CSV?

Chart2CSV is a tool that extracts numerical data from chart images. You give it a picture of a chart (like a screenshot or scan), and it gives you the data as a CSV file.

How do I extract data from a chart image?

python -m chart2csv.cli.main your_chart.png

This creates a CSV file with the extracted data points.

What chart types are supported?

  • Line charts
  • Scatter plots
  • Bar charts

Both linear and logarithmic scales are supported.

Is Chart2CSV free?

Yes. Chart2CSV is free and open source under the MIT license.

Does it work offline?

Yes. By default, Chart2CSV uses Tesseract OCR which runs completely offline. Your images never leave your computer.

What is Mistral mode?

Mistral is an AI service that can read text from images more accurately than traditional OCR. To use it:

  1. Get a free API key from https://console.mistral.ai/
  2. Set MISTRAL_API_KEY environment variable
  3. Add --use-mistral flag

How accurate is it?

Accuracy depends on image quality. For clean, high-resolution charts, accuracy is typically 95%+.

Use --overlay to verify results visually.

What if the results are wrong?

  1. Check the overlay image to see what was detected
  2. Try --use-mistral for better OCR
  3. Use --crop to focus on the chart area
  4. Use --calibrate for manual calibration

Can I process multiple charts at once?

Yes. Use batch mode:

python -m chart2csv.cli.main folder/ --batch --output-dir results/

What image formats are supported?

  • PNG
  • JPG/JPEG
  • WebP

How do I cite Chart2CSV?

@software{chart2csv,
  title = {Chart2CSV: Zero-Click AI Chart Data Extraction},
  author = {kiku-jw},
  year = {2025},
  url = {https://github.com/kiku-jw/Chart2CSV}
}

Where can I report bugs?

Open an issue on GitHub: https://github.com/kiku-jw/Chart2CSV/issues

Clone this wiki locally