Build deep learning parsing APIs for

Build, train and deploy state-of-the-art document parsing models powered by machine learning. No ML experience required.
MacBook mockup
Trusted by organizations
across industries

What can you build with Cradl AI?

Create powerful document parsers for a variety of applications and document formats.

High accuracy

Train your own document parser to ensure that
your data extraction is optimized for your documents.

Fully customizable

Extract information from any document in any
Latin-based language and any format.

Get started in 10 minutes

Train your first model on as little as 15 documents. Collect data, re-train and improve as you go.

Go beyond templates, rules and traditional OCR with deep learning

Since Cradl AI is pre-trained on millions of documents, you can train your own document parser using very little training data.
Automatically adapt to ever-changing document layouts
Set up a continuous training loop to adapt to new layouts automatically.
Any language, any document type
All Latin-based languages are supported, and numerical
values can be extracted from any language.
Dashboard mockupDeep learning heatmap

Predictions you can trust

Actionable confidence estimates for each fields means that you know when a prediction need to be validated by a human.
Straight-through processing
Reduce manual work with straight-through processing for
high-confidence predictions.
Stay in control
Accurate confidence estimates enable you to make your own
trade-off bewteen accuracy and automation.
OCR confidence threshold
AP automation software AI

Put humans in the loop
only when you need to

Our optional human-in-the-loop interface makes it easy to validate
and correct the data extracted from your documents.
Easy to use
The intuitive UI makes it easy for any reviewer to get up and running fast.
Train as you go
Your model learns from your human operators without the need for manual box annotation - increasing automation and accuracy over time.

All the features you need

Flexibility, speed and continuous improvements - all while staying secure and compliant.

Full flexibility

Extract any field from semi-structured
documentsin any Latin-based language.


Average processing time is
less than 2 seconds.

Continuous learning

Set up a seamless feedback loop to continuously improve your model.

Secure and compliant

We make sure your data stays safe and is stored and processed on servers in Ireland in full compliance with EU's GDPR directive.


Integrate anywhere

Integrate your document parsing models in any solution
with our easy-to-use REST API or use one of our SDKs.

Don’t just take our word for it

Companies across industries use Cradl AI to build powerful document automation.
We use Cradl to deliver the most frictionless invoice payment solution on the market for our customers. Our Cradl models are now automating more than 33 million keystrokes for our customers annually.
User 1
Afshin Khan
Product Manager @ DNB
Using Cradl, we created a powerful ML model that we integrated in our UiPath RPA workflow, able to automate expense control of more than 210 000 government employees.
User 2
Hallgeir Molde
Project manager @ DFØ
We're building the best software for accounting firms. Our customers have very high expectations, and with Cradl we can be confident that our document automation is best-in-class. We're super impressed with both accuracy and speed!
User 3
Anders Skjetne
CEO @ Centiga

Start building today

Get started for free, or book a demo to discuss your project.
Join our newsletter
Join our newsletter to stay up to date on releases.
We care about your data in our privacy policy.
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.