Openreading
Openreading is a document intelligence router that routes documents to other parsers with key techniques for managing failures, based on the author's team's enterprise document intelligence experience. It is open core and work-in-progress, with benchmarks and a managed version (durable compute, retries, parallel execution, intent understanding) planned for later.
What it addresses
Even with state-of-the-art OCR, SLMs, LLMs, and many parser services, most miss something in enterprise document intelligence, so failure handling has to be designed for.
Shared over time
Each entry reflects what was shared on that date. Details may have changed since then.
Openreading
Openreading is a document intelligence router that routes documents to other parsers with key techniques for managing failures, based on the author's team's enterprise document intelligence experience. It is open core and work-in-progress, with benchmarks and a managed version (durable compute, retries, parallel execution, intent understanding) planned for later.
Problem: Even with state-of-the-art OCR, SLMs, LLMs, and many parser services, most miss something in enterprise document intelligence, so failure handling has to be designed for.
Audience: Building document pipelines for sensitive industries like healthcare, insurance, and lending where existing parsing services miss things and failures need to be managed.
Posted by @c0mpute · Original comment ↗
“We are bringing these ideas out as a router to other parsers, but with some key takes that allow you to manage failures. "Failure is inevitable so route for it." It is called Openreading - https://openreading.ai/ and open core here: https://github.com/openreading-ai/openreading-core”