Features
Ease of use
READ offers a WOG portal that users can try out freely, simply by dragging and dropping files into the platform or uploading them in batches. It also provides APIs that seamlessly integrate into your existing systems, offering full end-to-end pipeline control and customisation.
Fast processing
READ achieves an average processing time of 1-2 seconds per document, minimising wait times and accelerating workflows.
High accuracy
READ achieves 95% accuracy on average, which is higher than that of similar commercially available solutions across diverse document types.
Wide ranging support
READ processes documents up to CONFIDENTIAL (CLOUD) / SENSITIVE HIGH, and even SECRET for certain modes of deployment. It also supports a range of document formats including PDF, JPG, PNG, and TIFF.
Elastic scalability
READ has an internally hosted large vision-language model, which allows quick scalability and adaptability to new use cases.
Roadmap
Scaling READ across agencies to support more document types and use cases
- READ team is actively engaging with multiple agencies, including ICA, MOM, DSTA, CPF, MSF, and IPOS, to validate their use cases and support system integration for full adoption.
- READ is concurrently building partnerships across WOG to establish itself as the preferred solution for document processing and understanding, driving seamless automation at scale.
Techstack
FastAPI, PyTorch, Transformers, OpenCV, PaddleOCR, OpenClip, Spacy, XGBoost, GliNER PyMuPDF, Ultralytics, AWS ECS, ECR, ALB, ASG, SageMaker, NAT, Lambda, EventBridge, DynamoDB, LMDeploy
Thanks for letting us know that this page is useful for you!
If you've got a moment, please tell us what we did right so that we can do more of it.
Did this page help you? - No
Thanks for letting us know that this page still needs work to be done.
If you've got a moment, please tell us how we can make this page better.
An AI-Powered Solution That Streamlines Document Automation for Faster, More Accurate Workflows