Mindee offers an API that lets you turn raw data in a paper document into structured data. The startup is coming out of stealth mode and has raised a $14 million Series A round led by GGV Capital.
Other investors in today’s funding round include Alven, Serena Capital, Bpifrance’s Digital Venture fund as well as several business angels, such as Algolia co-founder Nicolas Dessaigne and Datadog co-founder Alexis Lê-Quôc.
If you’re working on a product that involves saving paper receipts, scanning driver licenses or entering invoices in a system, chances are you’ve been looking at ways to automate that process as much as possible. Many companies rely on OCR services (optical character recognition) from Google or Amazon.
But turning text into raw digital text isn’t the most complicated step. After that, you want to turn raw digital text into structured data. This is much more complicated as companies often rely on humans to identify the right information and paste it in the right data field.
Essentially, you can process a photo through Mindee API and turn it into relevant data for your product without any manual data entry. Many industries could use this kind of API-based product, from expense management to procurement, accounting, loan applications and more.
Some of Mindee’s clients include Qonto, Payfit, Spendesk and Lucca. With the new funding, the startup plans to get more customers in Europe and expand to the U.S.
Behind the scenes, the company has trained its algorithms using machine learning on large data sets of documents. It is supposed to get better over time as it processes more documents. The company already processes millions of pages every month from its customers.
The startup charges its clients based on volume. Once you go over the free tier, you get charged per processed page. And because it’s a hosted API, integrating Mindee in your product doesn’t require a ton of changes in your product. You call an API and receive structured data in return.
What makes Mindee special is that the experience for the end user is much better. For instance, when you try to sign up to a service that requires your passport information, you usually have to enter your personal information, take a photo and wait. It’ll ping someone so that they can make sure it’s the same information in the registration system and on the photo.
If Mindee manages to capture data in near real-time and with a near perfect level of accuracy, you could get a much snappier experience. And that could lead to more customers for Mindee’s clients.