Footer

Your go-to destination for discovering trending open-source projects and uncovering the insights that matter

Legal

  • Terms and conditions
  • Privacy policy

More

  • Pricing
  • Advertise
  • onedump
  • gosnakego
  • Contact us

Social

  • X (Twitter)
  • GitHub
Trendshift owl
What they build speaks louder than words
© 2026 Trendshift, created by Julian Li with ❤️
SignalAdvertise
Trendshift
SignalAdvertise
Sign in
Trendshift owl

Discover trends that matter

DailyWeeklyMonthlyYearly
Live mentionsTopics
RepositoriesDevelopers
InsightsStats
Bifrost
Bifrost

baidu/Unlimited-OCR

Unlimited OCR Works: Welcome the Era of One-shot Long-horizon Parsing.

#Document parsing
Visit GitHub
Python
29.8k
3.3k
1 contributors
MIT License
AchievementsView all →
FeaturedAdvertise to 180K+ monthly visitors
Bifrost

Fastest Enterprise AI Gateway for Routing, Governing, and Securing all your AI Traffic

Momentum

How fast baidu/Unlimited-OCR is gaining stars, and where it lands on Trendshift and GitHub Trending

Activity and trending

Stars, forks and merged pull requests each day or month, lined up with the repository's rank on Trendshift Trending

Aug 31 – Oct 30, 2026

Stars +5,063 in 61 days

0200400600
568

Forks +651 in 61 days

050100
66

Merged PRs 0 in 61 days

01

Trendshift Trendingwhere the repository ranked each day it made the top 25

Top 25 in Python

198
Aug 31Sep 8Sep 16Sep 24Oct 2Oct 10Oct 18Oct 26
The left and right arrow keys step from day to day; Home and End go to the first and the last

Stars, forks, merged pull requests and Trendshift Trending rank, by day
DayStars gainedForks gainedMerged PRsPython
Mon, Aug 31, 2026+148+210Not trending
Tue, Sep 1, 2026+115+160Not trending
Wed, Sep 2, 2026+52+60Not trending
Thu, Sep 3, 2026+62+30Not trending
Fri, Sep 4, 2026+35+60Not trending
Sat, Sep 5, 2026+60+80Not trending
Sun, Sep 6, 2026+56+90Not trending
Mon, Sep 7, 2026+44+40Not trending
Tue, Sep 8, 2026+32+20Not trending
Wed, Sep 9, 2026+23+20Not trending
Thu, Sep 10, 2026+16+00Not trending
Fri, Sep 11, 2026+24+30Not trending
Sat, Sep 12, 2026+92+120Not trending
Sun, Sep 13, 2026+42+10Not trending
Mon, Sep 14, 2026+57+70Not trending
Tue, Sep 15, 2026+43+30Not trending
Wed, Sep 16, 2026+51+70Not trending
Thu, Sep 17, 2026+74+70Not trending
Fri, Sep 18, 2026+49+70Not trending
Sat, Sep 19, 2026+57+110Not trending
Sun, Sep 20, 2026+45+40Not trending
Mon, Sep 21, 2026+31+50Not trending
Tue, Sep 22, 2026+103+200Not trending
Wed, Sep 23, 2026+86+180Not trending
Thu, Sep 24, 2026+59+120Not trending
Fri, Sep 25, 2026+54+50Not trending
Sat, Sep 26, 2026+47+50Not trending
Sun, Sep 27, 2026+88+100Not trending
Mon, Sep 28, 2026+58+120Not trending
Tue, Sep 29, 2026+37+50Not trending
Wed, Sep 30, 2026+25+50Not trending
Thu, Oct 1, 2026+33+30Not trending
Fri, Oct 2, 2026+30+20Not trending
Sat, Oct 3, 2026+18+30Not trending
Sun, Oct 4, 2026+28+50Not trending
Mon, Oct 5, 2026+19+10Not trending
Tue, Oct 6, 2026+43+70Not trending
Wed, Oct 7, 2026+39+40Not trending
Thu, Oct 8, 2026+24+20Not trending
Fri, Oct 9, 2026+26+20Not trending
Sat, Oct 10, 2026+11+30Not trending
Sun, Oct 11, 2026+115+200Not trending
Mon, Oct 12, 2026+151+140Not trending
Tue, Oct 13, 2026+211+310#19
Wed, Oct 14, 2026+67+70Not trending
Thu, Oct 15, 2026+126+90Not trending
Fri, Oct 16, 2026+169+210Not trending
Sat, Oct 17, 2026+79+120Not trending
Sun, Oct 18, 2026+76+170Not trending
Mon, Oct 19, 2026+71+70Not trending
Tue, Oct 20, 2026+63+50Not trending
Wed, Oct 21, 2026+122+220Not trending
Thu, Oct 22, 2026+152+290Not trending
Fri, Oct 23, 2026+64+110Not trending
Sat, Oct 24, 2026+47+40Not trending
Sun, Oct 25, 2026+47+40Not trending
Mon, Oct 26, 2026+182+260Not trending
Tue, Oct 27, 2026+568+660#8
Wed, Oct 28, 2026+379+450Not trending
Thu, Oct 29, 2026+238+310Not trending
Fri, Oct 30, 2026+100+20Not trending
Total+5,063+6510
  • 1#1 on the list
  • 7Rank that day
  • Not trending
4

    • GitHub trending history

    Where the repository ranked on GitHub Trending, one dot for each day it was on a list

    Python

    Oct 13 – Oct 29, 2026Best #4

    #1#5#10#15#20#25
    Top 10
    Oct 13, 2026Oct 29, 2026
    The left and right arrow keys step from day to day; Home and End go to the first and the last

    Mentions and achievements

    Where people are talking about baidu/Unlimited-OCR, and every badge it has earned
    1. Oct 301 post

      • 整本几百页的书，一次请求就解析完——百度这个新模型我刚核对完，点开演示就能试 痛点我太熟了：论文、合同、教材拆成一页页喂，跨页的表格和上下文全断掉。这个 2 代换了个思路： ▌能领到什么：开源免费（MIT），不用部署也能先玩——仓库注明的 Hugging Face 新演示点开就能试解析 ▌一次解析整本：多页文档一次请求解析，跨页上下文不丢失；官方称在超长文档基准上达到 SOTA…

        Read full on X, posted by @404latitude

        @404latitude · x.com

    2. Oct 282 posts

      • The SOTA claims are on MPDocBench and MarathonDocBench, and they built the second one, so I'd treat that as self-reported for now. I've got a pile of scanned manuals I'd love to throw at it. What's…

        Read full on X, posted by @buildAgoat

        @buildAgoat · x.com

      • 超长文档解析，页数一多就断、表格就乱？ Unlimited-OCR 是百度做的超长文档解析：单次长程解析长文档，2 代已出。 对常处理长文档的人来说，单次解析比切段拼结果省事。 你会先拿它解析哪类文档，是论文还是财报？ #OCR #开源 @grok

        Read full on X, posted by @xiaoanlab

        @xiaoanlab · x.com

    3. Oct 272 posts3 posts

      • 兄弟们，做过三次 OCR 横评的我说句实话。（看视频） 文心 Unlimited OCR 2，可能是今年最被低估的OCR类的开源项目。 3623 页的《哈利·波特》七部曲，整本丢进去，不切分、不拼接，1 小时 30 分读到最后一页。 1222 页的《三体》，50 分钟，我实测了《平凡的世界》1600页，40分钟完成可下载的txt。 以前的 OCR 卡在哪？…

        Read full on X, posted by @berryxia

        @berryxia · x.com

      • 百度开源Unlimited-OCR 2无限文本 2-4G内存即可本地部署！ 先说结论： - PaddleOCR那个时代可以翻篇了 - 本地私有化部署相当友好，3-4G显存或同等内存就可高精度部署，2-3G内存可以流畅跑 1、解决什么场景，痛点？ 以前用传统OCR处理几十上百页的长PDF、深度研报、复杂财报，要么切图切到碎掉、排版全乱，要么多模态大模型直接 OOM 爆炸。…

        Read full on X, posted by @cfenghub

        @cfenghub · x.com

      • 最后还有个小彩蛋 Unlimited OCR Works 的名字，借用了日本知名游戏和动画《Fate/stay night》里的一个招式名：“无限剑制”（Unlimited Blade Works）。在作品中，这个招式能展开一个布满刀剑的空间，使用前还有一段经典台词。 第一代模型发布后，一位社区用户把这段台词改成了 OCR 版本： “像素为身，Token 为血 我已阅尽千余页文档……”…

        Read full on X, posted by @xiaohu

        @xiaohu · x.com

    4. 15 days earlier
    5. Oct 121 post

      • 百度的这个新OCR可以，Unlimited OCR，解决长文档OCR速度衰减和显存膨胀问题，几十页文档一次性匀速解析完 3B MoE（激活5亿参数），端到端文档解析，可本地部署 基于DeepSeek-OCR微调，核心是R-SWA，输出6000 token时比DeepSeek-OCR快约35% 传统LLM解码器的KV cache会随输出变长而不断膨胀，导致又吃显存又慢…

        Read full on X, posted by @aigclink

        @aigclink · x.com

    6. Oct 81 post

      • 百度这次把 OCR 做成开源项目，是不是有点猛？ 推荐一个开源项目 Unlimited-OCR，主打本地文档识别：图片可以批量丢，PDF 不用自己分页，长文档直接整份处理，还能自动高分辨率渲染后逐页 OCR。 模型约 3B 参数，支持多语言，兼容 Transformers、vLLM、Ollama 等部署方式。合同、论文、扫描书、古籍这种几十上百页的 PDF，用起来尤其省事。

        Read full on X, posted by @xiao92666

        @xiao92666 · x.com

    7. 21 days earlier
    8. Sep 171 post

      • • Supports single image, multi-page, and PDF parsing via PyMuPDF conversion • Runs with Transformers, vLLM, and SGLang for flexible deployment • Configurable gundam and base modes for different image…

        Read full on X, posted by @the_osps

        @the_osps · x.com

    9. 12 days earlier
    10. Sep 51 post

      • • Supports single image, multi-page, and PDF parsing with two configs: gundam and base • Inference via Transformers, vLLM, or SGLang with up to 32,768 context length • Available on Hugging Face,…

        Read full on X, posted by @the_osps

        @the_osps · x.com

    You might also like

    Repositories close to baidu/Unlimited-OCR
    FeaturedAdvertise to 180K+ monthly visitors
    Bifrost

    Fastest Enterprise AI Gateway for Routing, Governing, and Securing all your AI Traffic

    majcheradam/ocrbase

    TypeScript

    📄 PDF/IMG ->.MD/JSON Document OCR API for PaddleOCR and GLMOCR. Self-hostable.

    293

    deepseek-ai/DeepSeek-OCR-2

    Python

    Visual Causal Flow

    3.5k333

    deepdoctection/deepdoctection

    Python

    A Repo For Document AI

    3.3k205

    aiptimizer/TurboOCR

    C++

    TurboOCR, >200 img/s OmnidocBench. TensorRT FP16, PP-OCRv6, HTTP + gRPC

    1.1k120

    zai-org/GLM-OCR

    Python

    GLM-OCR: Accurate × Fast × Comprehensive

    7.6k701

    misbahsy/doc-router

    Rust

    A Document OCR Router to help route pages based on content.

    326

    datalab-to/chandra

    Python

    OCR model that handles complex tables, forms, handwriting with full layout.

    12.6k1.3k

    studio-dots-ai/dots.ocr

    Python

    Multilingual Document Layout Parsing in a Single Vision-Language Model

    9.2k822

    OpenBMB/MiniCPM-OCR

    Python

    An open-source OCR solution for fast, accurate document parsing with correctness assessment.

    59627