▌ GitHub radar

Baidu's One-Shot Long Document OCR

Baidu released an open-source OCR engine that parses entire multi-page documents and PDFs in a single pass, not just individual screenshots.

01baidu/Unlimited-OCR 379Python

Baidu's Unlimited-OCR extends DeepSeek-OCR to handle multi-page documents and PDFs in a single model call, rather than forcing you to feed pages one by one. It supports batch inference, streaming API responses, and integrates with standard inference frameworks. Nearly 400 stars in a few days after release.

Why a vibe-coder should care

If you've ever tried to extract text from a long PDF and ended up splitting it page by page, this solves exactly that: one request, the entire document, structured output ready to use.

Open on GitHub →