Skip to content

Latest commit

 

History

History
14 lines (11 loc) · 804 Bytes

README.md

File metadata and controls

14 lines (11 loc) · 804 Bytes

receiptchat

Use generative AI to extract information from receipts

Try it out

  • Make sure tesseract is installed (on mac, can be installed using homebrew or macports)
  • Clone this repo, make a python 3.9+ environment and install the requirements
  • Set up your Google Drive credentials and put them on a file called "gdrive_credentials.json". Store that in the top level directory
  • Set up an OpenAI API key and store it in a .env file in the top level directory
  • Put some receipts in your connected Google Drive folder
  • To run vision extraction (generate the examples), run driver.main_generate_examples(). By default only 10 files will be parsed
  • To run text extraction (i.e. with gpt-3.5-turbo learning from examples), run driver.main()

Be aware of your API costs!