Still cleaning PDFs in the agent era?
Shanghai AI Lab’s Sciverse: 30M+ AI-ready full texts from 10+ sources incl. arXiv.
Free. Parsed. Smart search. T+1 updates. API/Skills for batch agent calls.
Do science. Skip the data prep. sciverse.opendatalab.org.cn/
I just published 30.71 Million Full Texts, Free for Research: The MinerU Team Brings Scientific Knowledge Straight… medium.com/@opendatalab.feed…
MinerU 4.0: Model + Skill upgrades!
Model: More formats, 4 parsing modes & source locators for agents to revisit originals.
Skill: Auto environment checks, on-demand setup, local-first parsing. Uploads require consent.
Try it: github.com/opendatalab/Miner…
🤖 Made with AI
Congratulations to Prof. Bowen Zhou on being named to the 2026 #TIME100AI list!
we share his commitment to advancing AI for Science—building open data and tools that help AI deepen scientific inquiry and expand the boundaries of discovery.
Read more: time.com/collection/time100-…
MinerU.Chem extracts chemistry papers into AI-ready data.
Molecules, reactions, identifiers -> SMILES/MolFile, with source traceability.
93.02% SMILES exact match on MolRecBench-Wild.
Try it: mineru.net/OpenSourceTools/E…
🎉 Introducing NanaDraw – your AI-powered scientific diagram maker!
Generate editable, publication-ready flowcharts, architecture diagrams, and illustrations in 5 minutes with simple text prompts.
✅ 4 smart modes: Auto/Draft/Generate/Assemble
shannon.opendatalab.com/nana…
🚀 Now MinerU 3.1.0
Full integration with MinerU2.5-Pro
✅ Apache 2.0-based license (more business-friendly)
✅ Native DOCX/PPTX/XLSX parsing
✅ 8GB RAM handles 10k+ pages
✅ Multi-machine/multi-card, high concurrency
From a tool to industrial-grade parsing infrastructure
#MinerU
● GitHub: github.com/opendatalab/Miner…
● HuggingFace: huggingface.co/spaces/openda…
● ModelScope: modelscope.cn/models/OpenDat…
🚀 Introducing MinerU Document Explorer!
Inspired by Karpathy’s LLM Wiki vision — build your dynamic, self-maintaining AI knowledge base beyond RAG.
Parse PDFs/Word/PPT, auto-generate linked Wiki, precise extraction & full traceability.
Lightweight, local, agent-native.
👉 GitHub: github.com/opendatalab/Miner…
🚀MinerU2.5-Pro!
Same 1.2B architecture, NO structural changes.
65.5M pages, diversity-aware sampling, cross-model verification, render-then-verify correction.
Scores 95.69 on OmniDocBench v1.6, outperforming larger models.
Built for real-world use
Now open-sourced!
#MinerU
OpenDataLab retweeted
结果就是在 OmniDocBench v1.5 上, 0.9B 参数 + 2.5K 视觉 token 就拿下了 92.62 的综合分, 文本/公式/表格/阅读顺序四项全部拿到了 SOTA 的成绩. 超过了 Gemini-2.5 Pro (88.03), Qwen2.5-VL-72B (87.02), MinerU2.5 (90.67). 推理速度也比其他的模型快, 而且显存占用更少.
论文证明了观点：在大模型时代, 精巧的工程设计 + 高质量数据可以让小模型达到甚至超越大模型的水平. 论文对整个 OCR 领域都有很大的启发意义👍.
Struggling with messy receipts & invoices? 🧾
#MinerU Skills delivers zero-code receipt parsing.
Automatically locate, split & extract key data—amounts, dates, items—with high accuracy.
Structured output ready for storage & reconciliation.
Full workflow tutorial video now live!
Tired of messy PDF outputs? 📄
#MinerU Skills let you process papers with zero code.
Parse layouts, formulas, tables & OCR in one click.
Batch 50+ papers to clean Markdown, keep LaTeX & tables perfectly.
Connect to RAG/knowledge base effortlessly.
Full tutorial video out now!
🚀#OpenDataLab’s AI-ready database #Sciverse Launched!
Powering #AGI4S with a 3-layer system (Sci-Base/Sci-Align/Sci-Evo).
✅25M+ parsed literatures, 600B+ high-quality tokens via ＃MinerU
✅18M+ protein sequences, 6M+ chemical reactions
👉Explore: sciverse.opendatalab.com
🚀New from #OpenDataLab: MinerU-Diffusion!
We redefine document OCR as inverse rendering via diffusion decoding, replacing slow autoregressive generation.
✅ Up to 5.1× faster inference
✅ Stronger visual structure modeling
✅ Stable in challenging scenarios
Try it & star us!
Free open-source PDF parser MinerU now offers Skills, MCP Server, dual-mode API, cross-platform CLI/SDK, and RAG plugins.
One sentence lets AI read PDFs easily!
🦞MinerU skills
clawhub.ai/MinerU-Extract/mi…
👉Dev-ecosystem Usage Guide
mineru.net/ecosystem
🚀Big Update! MinerU has adapted to 10+ computing power platforms
💯99% accuracy in capturing PDF/web elements
💪OmniDocBench adopted by Gemini3/DeepSeek as authoritativee benchmark
👉Explore: github.com/opendatalab/miner…;
mineru.net/OpenSourceTools/E…
🏆MDIC: mineru.net/MDIC2026
