Skip to content

A receipt reader that analyses a photograph of a printed or handwritten receipt and extracts individual items, quantities, and prices. A digital version is then generated and automatically attached to the transaction in the Monzo app.

License

Notifications You must be signed in to change notification settings

kimbethstonehouse/monzo-receipt-reader

 
 

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

53 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Monzo Receipt Reader

Results

This project won the Monzo Challenge at Hack Cambridge 4D in January 2019.

What is it?

A receipt reader that analyses a photograph of a printed or handwritten receipt and extracts individual items, quantities and prices. A digital version is then generated and automatically attached to the transaction in the Monzo app.

How does it work?

Click on the relevant transaction in the Monzo app, click add receipt and then upload a photograph of the receipt.

Demo

Step 1

Paper

Step 2

Monzo

Step 3

Digital

Technologies

We used Microsoft Azure's Vision API to get a raw text dump file. We then took the raw text dump and tried to analyse it line by line to find structure and patterns. We then extract each item (quantities, singular item prices, total item prices) and feed it into the Monzo API.

The Monzo API refreshes at an interval to look for new attachments for transactions, runs it through the API and tabulates it into the receipt. This is done all natively and automatically through the Monzo app.

We also used Microsoft's Azure Vision API to add support for handwritten receipts, in case you're shopping at an old school shop.

About

A receipt reader that analyses a photograph of a printed or handwritten receipt and extracts individual items, quantities, and prices. A digital version is then generated and automatically attached to the transaction in the Monzo app.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages

  • Python 100.0%