Mindee is a developer platform that enables you to extract key information from any type of document with your code. Our mission is to help developers automating paperwork in their applications in real-time, with human-level accuracy.
Extracting a set of specific information from documents is also called Key information extraction. The goal is to define a list of fields you want to extract from documents, so you are able to automate a workflow based on this data when a new document comes up.
For many accounting use cases such as accounts payable automation, you need extract key information from invoices such as:
- Total amount
- Invoice date
- Due date
- Merchant name
- Invoice number
You can extract automatically those different fields using our Invoice OCR API.
Each of our off-the-shelf APIs were trained using hundreds of thousands of documents to provide you with the most robust parsing solution. We frequently release more off-the-shelf APIs such as our Receipt OCR API or Passport OCR API. Feel free to ask us our roadmap to see what comes next.
In case you need an API that was not already released in our catalog, you can still build your own in minutes.
Using our API Builder, you can define your own list of key information and train an API to extract those fields by annotating a few dozens of examples. Our users have already built and deployed for productions many different document parsing APIs, you can find a non-exhaustive list here.
Sometimes, you need to classify documents automatically in your code. One reason can be that your users upload several different types of documents at once, and you need to identify which is which. Perhaps they upload a single pdf including many different documents. It can be very tricky to automate this depending on your use case.
Using our API builder, you can train a powerful document classification API in minutes, using your own class definitions.
Updated about a month ago