document OCR / document understanding
GLM-OCR is a multimodal OCR model for complex document understanding, combining a GLM-V encoder-decoder with a layout analysis and parallel recognition pipeline for high-accuracy text extraction.
Same problem, same approach. Swapping one for another is a config change, not a rewrite.
Choose for a comparable OCR model with strong layout parsing for complex tables/forms/handwriting; evaluate on your own benchmarks.
OCR model that handles complex tables, forms, handwriting with full layout.
Choose for a research-focused document image parser using a different anchor-prompting approach.
The official repo for “Dolphin: Document Image Parsing via Heterogeneous Anchor Prompting”, ACL, 2025.
Choose for a comprehensive Python toolkit for PDF content extraction including OCR, layout, formula, and table.
A Comprehensive Toolkit for High-Quality PDF Content Extraction
Use for a lightweight Python tool converting images to Markdown with small models, ideal for resource-constrained setups.
An Open-Source Python3 tool with SMALL models for recognizing layouts, tables, math formulas (LaTeX), and text in images, converting them into Markdown format. A free alternative to Mathpix, empowering seamless conversion of visual content into text-based representations. 80+ languages are supported.
Choose for broader language support (90+) and an actively maintained OCR pipeline with full layout analysis.
OCR, layout analysis, reading order, table recognition in 90+ languages
Use for a modular Python library for text detection and recognition in your own pipeline.
docTR (Document Text Recognition) - a seamless, high-performing & accessible library for OCR-related tasks powered by Deep Learning. Ongoing development and maintenance by t2k.
Choose for a production-ready OCR toolkit with 100+ languages and extensive deployment options.
Turn any PDF or image document into structured data for your AI. A powerful, lightweight OCR toolkit that bridges the gap between images/PDFs and LLMs. Supports 100+ languages.
Solves the same problem with a different architecture or at a different layer. Expect to rewrite the integration.
Use when you prefer a TypeScript library that leverages general vision LLMs for OCR instead of deploying a dedicated model.
OCR & Document Extraction using vision models
Use when you need a fast Rust-based document parser with structured output, rather than a model-based OCR.
A fast, helpful, and open-source document parser
Use for OCR-free document understanding with a Transformer, avoiding traditional OCR pipelines.
Official Implementation of OCR-free Document Understanding Transformer (Donut) and Synthetic Document Generator (SynthDoG), ECCV 2022
An earlier or more limited way to do the job. Still the right call when you need something small, proven, or CPU-only.
Use for specialized image-to-LaTeX formula recognition, not full document OCR.
TexTeller can convert image to latex formulas (image2latex, latex OCR) with higher accuracy and exhibits superior generalization ability, enabling it to cover most usage scenarios.
Use for a simple Tesseract wrapper when you need lightweight local OCR without deep learning.
:clipboard: Python wrapper to grab text from images and save as text files using Tesseract Engine
Use for OCR from Go via Tesseract, a lightweight traditional engine.
Go package for OCR (Optical Character Recognition), by using Tesseract C++ library
Use for Japanese manga text OCR specifically, not general documents.
Optical character recognition for Japanese text, with the main focus being Japanese manga
Use for converting equation images to LaTeX code, a focused subset.
pix2tex: Using a ViT to convert images of equations into LaTeX code.
Use for handwritten text recognition with a legacy TensorFlow implementation.
Handwritten Text Recognition (HTR) system implemented with TensorFlow.
Does the same job, but ships as an app. Useful to a person, not swappable into a codebase.
Use as a CLI tool for one-off plaintext extraction without GPU or cloud.
Snap any image, screenshot, or webpage into plaintext. No GPU. No cloud. One command.
Use as an end-user PDF conversion tool for scanned books.
PDF craft can convert PDF files into various other formats. This project will focus on processing PDF files of scanned books.
An archived Android app, not a library to integrate.
Experimental optical character recognition app
Use for an end-user math recognition/editing workspace with handwriting input.
Formula Recognition & Office Editing Math Workspace | Handwriting & PDF to LaTeX/Markdown, powered by MathCraft OCR & MathLive
These projects were analysed and named GLM-OCR among their alternatives. The relationship is not symmetric — how GLM-OCR rates them is a separate judgement, made when GLM-OCR is analysed in its own right.
calls GLM-OCR “Choose when you want a unified multimodal OCR model that integrates with the GLM ecosystem, if you can use that stack.”
Ready-to-use OCR with 80+ supported languages and all popular writing scripts including Latin, Chinese, Arabic, Devanagari, Cyrillic and etc.
calls GLM-OCR “VLM OCR model from a different lab; choose for its own speed/accuracy, but verify output format.”
Unlimited OCR Works: Welcome the Era of One-shot Long-horizon Parsing.
calls GLM-OCR “choose as a deep-learning OCR model alternative with similar speed and accuracy goals.”
Contexts Optical Compression
calls GLM-OCR “drop-in peer from Zhipu AI with similar OCR/layout/table coverage.”
OCR, layout analysis, reading order, table recognition in 90+ languages
calls GLM-OCR “choose as an alternative Apache-2.0 OCR model with its own performance tradeoffs.”
OCR model that handles complex tables, forms, handwriting with full layout.
calls GLM-OCR “choose when you want the GLM-family OCR VLM with Apache-2.0 weights and fast inference.”
Multilingual Document Layout Parsing in a Single Vision-Language Model
calls GLM-OCR “when you want an Apache-2.0 VLM-based OCR/document parser as a direct model alternative.”
The official repo for “Dolphin: Document Image Parsing via Heterogeneous Anchor Prompting”, ACL, 2025.
calls GLM-OCR “end-to-end OCR model with similar capabilities, choose as a direct alternative with Apache-2.0 license”
Official code implementation of General OCR Theory: Towards OCR-2.0 via a Unified End-to-end Model
calls GLM-OCR “Choose when you need an end-to-end vision-language OCR model that reads images without separately detecting text boxes.”
📄 Awesome OCR multiple programing languages toolkits based on ONNX Runtime, OpenVINO, MNN, PaddlePaddle, TensorRT and PyTorch.
calls GLM-OCR “OCR model from Zhipu; choose for fast/accurate recognition”
PandaOCR - 多功能OCR图文识别+翻译+朗读+弹窗+公式+表格+图床+搜图+二维码
calls GLM-OCR “Choose if you need an end-to-end OCR model without a separate LLM correction step.”
Enhances Tesseract OCR output using LLMs (local or API) for error correction, smart chunking, and markdown formatting of scanned PDFs
calls GLM-OCR “A transformer-based OCR model with a different architecture, suitable when you want a single powerful model rather than a full toolkit.”
OpenOCR: An Open-Source Toolkit for General-OCR Research and Applications, integrates a unified training and evaluation benchmark, commercial-grade OCR and Document Parsing systems, and faithful reproductions of the core implementations from a wide range of academic papers.