openparser

If you've ever extracted json data from files using llms, you know the pain in deciding if you can trust it or not. We found a solution that works well for our users: we track the full lineage of every extracted filed from the source document to the structured output. Our approach takes into consideration ocr (parsing), extraction and transformations steps. Results come with confidences attached, so you know which fields need a double check and approval from a human. And this human-in-the-loop mechanism, where humans only check low confidence results,. can be used effectively to speed up existing backoffice operations that deal with document workflows. You can try it right now at openparser.dev , or read this (https://www.openparser.dev/blog/how-openparser-uses-lineage) on how we built it and why one-shotting an llm to extract data from

We build document AI workflows, and OCR is finicky. At scale, it becomes hard to debug errors and compare confidence scores across providers. New OCR models appear all the time, so we wanted one interface that lets us swap models in and out without rewriting around each provider. So far, we’ve integrated nine models from PaddleOCR, Mistral, Azure, Google and AWS Textract. We run them at cost, with no markup. We also created and open-sourced a shared schema, openparser@1, plus adapters that normalize each provider’s response. You can switch OCR models without changing the rest of your application: https://github.com/eigenpal/openparser I’d appreciate feedback on the schema and which models or providers we should add next.

I built OpenParser because I wanted PaddleOCR-VL-1.6 behind an endpoint I could actually use. Now there are open-sourced CLI and TypeScript/Python SDKs: ``` npm install -g @openparser/cli openparser auth login openparser parse sync document.pdf ``` That is basically it. It doesnt get easier and cheaper that this. Give it a PDF and get back the text and document structure. Repo: https://github.com/eigenpal/openparser I would love feedback from anyone willing to try the CLI and tell me what is still annoying.