Skip to main content

What is Cradl?

Cradl is a platform where developers can utilize deep learning to capture data from their documents. Use your own data to train advanced machine learning models through a simple API.

Design Philosophy

Be the easiest way for developers to build customized data capture APIs powered by deep learning models.

Who is Cradl for?

Cradl is built for developers and works best if:

  • You are processing high volumes of semi-structured documents like invoices, receipts, forms, etc.
  • You require high-accuracy data capture.
  • You require low-latency data capture.

Key concepts

Cradl is centered around the following key concepts which you will frequently encounter in these docs:

  1. Models. A model is a machine learning model that can be trained to solve a specific document data extraction task, like capturing payment info from invoices.
  2. Documents. A semi-structured document like an invoice, receipt, ID document, etc. in a PDF or image format. Any language and any layout.
  3. Datasets. A dataset consists of many (typically thousands) of documents that are used to train a specific model.
  4. Data bundles. A data bundle is a collection of one or more datasets that has been prepared for Model training.
  5. Predictions. When your model has finished training using your Data Bundle it can start capturing data from new and previously unseen documents.

Getting started

Ready to get started with Cradl? Sign up for a free account, and start building.

How Cradl works illustration

Check out one of our quickstart guides, or read more about concepts to get a feel for how the platform works.