multilingual-ocr

A self-hosted PDF OCR API that converts scanned documents to markdown. Powered by PaddleOCR-VL, runs on GPU via Docker.

Auditable page-level run ledger for OCR and document extraction

🔍 廖工AI设计实战出品 | LiaoGong-OCR — easyocr+tesseract双引擎OCR，15条预处理链，手机拍屏数字识别87%准确率 | Dual-engine OCR with 15 benchmarked preprocessing chains, 87% phone-photo digit accuracy

OpenAI-compatible, vLLM-served OCR API for the Surya-OCR-2 model — multilingual document OCR (layout + text recognition) with request batching, a local CLI, and Docker packaging.

A minimal Windows desktop GUI for OCRmyPDF with multilingual OCR, batch processing, progress tracking, safe outputs, and a monochrome PySide6 interface.

A complete engineering case study in building an event-driven OCR pipeline: system design, distributed workers, hybrid AI model routing, and production observability - not just an OCR API wrapper.

Multilingual scene-text recognition for Indian languages using CLIP, EasyOCR, translation, text-to-speech, and Streamlit.

Scans packaged-commodity labels and checks every declaration against the Legal Metrology (Packaged Commodities) Rules, 2011 — itemised, clause-cited compliance reports from a single photograph. Multilingual OCR (English/Hindi/Gujarati), placement and numeral-height checks, PDF/DOCX export, enforcement dashboard. SIH 2026 · PS SIH26034