SaaS, also called software as a service, has been one of the most popular forms of software business models over the last few years. Klippa has been in the SaaS industry for over 5 years, with multiple out of the box solutions for invoice & receipt processing, invoice & receipt approvals and expense management. All of these services are built on top of our core competence: OCR. At Klippa, we automate document data extraction for a living. In this blog we will be explaining more about the Klippa receipt parsing API. An API as a service solution that converts pictures of receipts into structured data, that anyone can implement in their own application within hours. Perfect for data analysts, banks, accounting software and loyalty players.
How does receipt OCR work?
Parsing a receipt consists of a few steps. The total process takes less than 3 seconds with our receipt API. The first step is of course for someone to make a picture of the receipt. That picture can be send to our API. In the API we first check the document quality, if that is good enough we convert the receipt with OCR to a raw text file. You can compare this with a notepad text file. From there we check the document language and country of origin using machine learning. When that information is clear, we start identifying potentially important data points like dates and amounts. When we have listed all potentially important information, a decision engine chooses the right values, adds categories and converts the data to a JSON output. This JSON output is then sent back to the customer. And all of that within just 3 seconds!
So how does that look?
The image below gives you a simplified example of the input and output of a random receipt via our OCR API for receipts as a service. In this case it’s a receipt from a Dutch restaurant (Klippa is originally Dutch), but it can be a receipt from any country in any sector.
What receipt fields can Klippa extract with OCR?
The short answer to this question would be anything you like! Klippa has a very flexible OCR engine that is customisable for each customer. Out of the box we have the fields listed below, but on request we can add additional fields for your use case. Input can be files such as JPG, PNG and PDF, the default output is a JSON file. These are the default fields:
- Receipt number
- Date of purchase
- Merchant name
- Bank account number
- VAT number
- Phone number
- Tax Identification Number
- Chamber of Commerce number
- VAT amount per percentage
- VAT percentages
- Total VAT amount
- Total amount, VAT excluded
- Total amount, VAT included
- Payment method
- Card number
- Line-items names (products bought)
- Line-items prices
- Line-items category
What languages and countries does Klippa support?
Basically the answer here is the same, our receipt parsing API is flexible, so it can support any language or country. Out of the box Klippa is optimised for languages like English, Dutch, French, Spanish, Italian, Portuguese, German and more. Do you have a use case for a different language? Just reach out to us. We love a challenge!
Why is Klippa the best receipt parsing API?
The answer to this question might be more simple than you expect. And to find the answer, we have to go back in the history of OCR development. Because around 10 years back, companies started using OCR to try and automate invoice processing. They choose invoices because it’s one of the largest document streams worldwide and invoices are relatively structured documents. The more structure a document has, the easier it is to provide an OCR as a service solution. The software that most companies made where therefore being optimised for invoices. In the years following, the world has been digitised more and more. Therefore the request for other OCR solutions came up, one of them was receipt processing. Our competitors have of course also moved into this space, but they did so with their invoice-based solutions. If you compare invoices to receipts, you will see that receipts are much more chaotic and structured differently. Invoice-based solutions will therefore perform poorly on receipts.
So what did we do differently? Klippa actually started as a receipt scanning app. Back in 2014, one of the first in the app stores. Our receipt parsing API has therefore been optimised specifically for receipts. That means that we get a much higher accuracy, more detailed fields, but also a much better speed. Klippa is 100% cloud based so you will also benefit from our continuous improvement approach. Every two weeks an improved version of our receipt API goes live.
What is the next step?
We love to show our software in action. That’s why we offer free online demo’s to showcase what our software can do and have fast it is. You can schedule a demo below if you are interested. In the demo you can ask any questions you have and one of our product experts will help you out. If all the information in the demo is clear, the next step could be testing our receipt OCR parsing API using your own API key. Usually clients are up and running within 2 hours! Ready to get started?