What’s powering Mango?

Powered by artificial intelligence, Mango uses technologies like OCR, computer vision, and facial recognition to help you extract, organize, search, and analyze invoices, contracts, and images.

How does text extraction work?

Mango uses a combination of OCR and statistical modeling algorithms to interpret documents. OCR stands for optical character recognition, which describes the process of extracting text content from an image. OCR is just one of the technologies that enable Mango to extract accurate data from documents.

Whether it’s an image of a handwritten note or a scan of an invoice, Mango will be able to find your data and convert it to usable information for you. By using Mango you’ll no longer need to retype text saving you time.

Hello in cursive, hello with green boxes around each letter, hello with green boxes split out

First, advanced OCR algorithms scan your document for visual patterns to identify text within your document, image, or PDF.

Hello abcdefghijklmnopqrstuvwxyz

Second, Mango's machine learning algorithms analyze phrases of text to calculate several features, including relative position, textual patterns, and other semantic signals.

Hello small
Hello large

Third, after identifying and learning the best features, Mango will present you with recognizable and usable text in each field you define. Mango is unique in that it can map from raw text to the fields you define, without writing a single line of code!

How does image tagging work?

female with flower crown with dots and lines on her face

First, Mango uses a series of deep learning models to calculate several features about your image, including faces, objects, colors, and other visual signals.

4 photos: one larger photo of a female with a flower crown, three smaller images of two different people and one thats the same person as the larger photo

Second, these features are compared to your unique and private database of images and features to determine which labels are most relevant to your image

female with a flower crown with her face tagged with her name Alice Allen

Third, Mango displays this metadata in the user interface with an indication of confidence. Every time you confirm, correct, or add information, Mango learns a little bit more about your data and improves existing and future labels.

What makes Mango unique?

No code

Mango  is unique in that it can map from raw text to the fields you define, without writing a single line of code!

Continuous learning

Each new picture and new label improves Mango’s ability to accurately identify people in your corpus of images.