{"full_name":"HKUDS/RAG-Anything","owner":"HKUDS","name":"RAG-Anything","shard":"1","url":"https://github.com/HKUDS/RAG-Anything","description":"\"RAG-Anything: All-in-One RAG Framework\"","homepage":"http://arxiv.org/abs/2510.12323","topics":["multi-modal-rag","retrieval-augmented-generation"],"tag_pages":["retrieval-augmented-generation"],"language":"Python","license":"MIT","avatar":"https://avatars.githubusercontent.com/u/118165258?v=4","created_at":"2025-06-06T06:47:29Z","pushed_at":"2026-11-16T15:36:03Z","metrics":{"stars":24594,"forks":2881,"watchers":116,"issues":172,"prs":295,"releases":24,"issues_open":78,"issues_closed":94,"prs_open":6,"prs_closed":76,"prs_merged":213,"discussions":24},"growth":{},"history":{"dates":["2026-11-18"],"stars":[24594],"forks":[2881],"watchers":[116],"issues_open":[78],"prs_open":[6]},"boards":{},"same_owner":[{"full_name":"HKUDS/AI-Researcher","name":"AI-Researcher","shard":"1","boards":{}},{"full_name":"HKUDS/AI-Trader","name":"AI-Trader","shard":"1","boards":{}},{"full_name":"HKUDS/AutoAgent","name":"AutoAgent","shard":"1","boards":{}},{"full_name":"HKUDS/AutoSwarm","name":"AutoSwarm","shard":"1","boards":{"daily_rate":116}},{"full_name":"HKUDS/CLI-Anything","name":"CLI-Anything","shard":"1","boards":{"daily":56,"monthly":83,"daily_forks":120,"daily_watchers":96}},{"full_name":"HKUDS/ClawTeam","name":"ClawTeam","shard":"1","boards":{}},{"full_name":"HKUDS/DeepCode","name":"DeepCode","shard":"1","boards":{}},{"full_name":"HKUDS/DeepTutor","name":"DeepTutor","shard":"1","boards":{"daily":77,"weekly":82,"monthly":20,"monthly_rate":86,"daily_forks":134}},{"full_name":"HKUDS/LightRAG","name":"LightRAG","shard":"1","boards":{}},{"full_name":"HKUDS/OpenHarness","name":"OpenHarness","shard":"1","boards":{}},{"full_name":"HKUDS/Vibe-Trading","name":"Vibe-Trading","shard":"1","boards":{"monthly":191}},{"full_name":"HKUDS/nanobot","name":"nanobot","shard":"1","boards":{}}],"issues":[{"number":70,"title":"[Question]:Image is not analyzed properly","comments":10,"created_at":"2025-07-29T15:20:59Z"},{"number":49,"title":"[Question]: Not able to Process documents, Document Processing stuck","comments":9,"created_at":"2025-07-09T08:40:37Z"},{"number":235,"title":"[Question]:Consultation on the 63.4% accuracy on DocBench","comments":8,"created_at":"2026-04-02T06:43:56Z"},{"number":69,"title":"[Question]: raganything_example.py 改写为ollama下的模型时出现的问题","comments":7,"created_at":"2025-07-29T08:15:34Z"},{"number":146,"title":"[Question]:能够支持本地模型了吗","comments":6,"created_at":"2025-10-24T07:29:31Z"},{"number":50,"title":"[Question]:ImportError: cannot import name 'LightRAG' from 'lightrag' (unknown location)","comments":5,"created_at":"2025-07-10T06:31:06Z"},{"number":48,"title":"[Question]: Is it normal for documents to take a while to process?","comments":5,"created_at":"2025-07-08T15:14:04Z"},{"number":26,"title":"[Question]: How to Customize the Prompt for Knowledge Graph Generation in RAGAnything?","comments":5,"created_at":"2025-07-02T20:00:25Z"}],"discussions":[{"number":492,"title":"Chopper: a small open model that cuts only the needed sentences out of retrieved text","comments":0,"category":"Show and tell","created_at":"2026-11-16T09:09:37Z"},{"number":378,"title":"A hard dollar limit on one RAG-Anything ingestion job (text + vision calls)","comments":0,"category":"Show and tell","created_at":"2026-10-03T17:30:03Z"},{"number":245,"title":"File Metadata and Filtering","comments":2,"category":"Q&A","created_at":"2026-04-19T19:22:51Z"},{"number":229,"title":"Example code but using ollama","comments":2,"category":"Q&A","created_at":"2026-03-23T07:29:12Z"},{"number":194,"title":"Are there any plans to integrate RAG-Anything into LightRAG server?","comments":0,"category":"General","created_at":"2026-02-08T16:45:29Z"},{"number":193,"title":"Need a help!!!!  Why I always get this connection problem","comments":1,"category":"Q&A","created_at":"2026-01-23T10:08:44Z"},{"number":192,"title":"Feature Discussion: Audio Transcription Support","comments":1,"category":"Ideas","created_at":"2026-01-20T14:55:06Z"},{"number":179,"title":"加载已有的RAG库时报错。","comments":0,"category":"General","created_at":"2025-12-23T01:57:52Z"}],"latest_release":{"name":"v1.4.4","tag":"v1.4.4","published_at":"2026-10-28T07:03:46Z","body":"## MinerU compatibility\n\nRAG-Anything now requires **MinerU >=3.4.1,<4 on Python 3.10–3.13**. MinerU 4 replaced the CLI used by this integration. The parser now rejects incompatible versions at initialization and includes the actual command output and repair advice in errors (#449).\n\nOn **Python 3.14**, install RAG-Anything with Docling and select `parser=\"docling\"`, or use Python 3.13 for MinerU. MinerU 3.x cannot be installed on Python 3.14, so this release excludes that dependency there. The configured parser is checked at initialization, including operations that insert pre-parsed content.\n\nFor existing Python 3.10–3.13 environments that picked up MinerU 4:\n\n```bash\npip install -U raganything\npip uninstall -y docvortex mineru-llama-cpp\nmineru --version\npip check\n```\n\nThe upgrade selects MinerU 3.x. Removing `docvortex` resolves a leftover `pypdfium2` dependency conflict. A fresh virtual environment also avoids these leftovers. **Keep `mineru-magika` in an upgraded environment**: its installed files overlap with `magika`, so uninstalling it can delete the module MinerU 3.x needs. If that already happened, reinstall the installed `magika` version, for example `pip install --force-reinstall --no-deps magika==1.0.3` for the validated environment.\n\nMinimal Debian/Ubuntu images need the OpenCV system libraries:\n\n```bash\napt-get update && apt-get install -y --no-install-recommends libgl1 libglib2.0-0\n```\n\n## Security: standalone Markdown-to-PDF conversion\n\nThe WeasyPrint backend of `raganything.enhanced_markdown` now loads only `data:` resources and local images, CSS and fonts within the Markdown file's directory and its subdirectories. Resolved paths and extensions are checked, including symlinks. Remote resources, files outside that directory and other attachment types are refused (#446).\n\nThis changes the standalone PDF conversion utility; document ingestion uses the direct Markdown parser. For string conversion, supply `base_url` to allow local resources. **Pandoc has its own resource loading and is unchanged.** Since `method=\"auto\"` chooses Pandoc when installed, explicitly select `method=\"weasyprint\"` for these restrictions.\n\nIf you converted untrusted Markdown with earlier WeasyPrint configurations, check generated PDFs before sharing them: referenced local files may have been embedded or attached.\n\n## Fixes\n\n- **Ingestion:** re-inserting an already-processed multimodal document returns to `PROCESSED`; README-style equation items retain their LaTeX formula and description; captions and footnotes reach model prompts as readable text (#446).\n- **VLM queries:** retrieved image paths preserve spaces, uppercase extensions and full filenames while retaining the existing image-path validation (#445).\n- **Tables and equations:** preserve every table cell's column; handle Word function names, limits, tracked deletions and hidden radical degrees correctly (#423, #422, #421, #443).\n- **Text parsing:** preserve Japanese kana joins without deleting ordinary word spaces; skip null caption items; give regex-fallback entities useful names (#441, #444, #416).\n- **Batch processing:** start per-file timeout accounting when the worker starts, report timeouts as failures, and reparse a file if its output directory was removed (#434, #431).\n- **Office conversion:** isolate generated PDFs by source path so same-stem documents cannot overwrite one another (#437).\n- **Parse cache:** include result-affecting Docling options, PaddleOCR `cls` and MinerU `vlm_url`; the URL is hashed in cache metadata (#432, #446).\n- **Configuration:** read environment defaults when each `RAGAnythingConfig` is created (#428).\n- **Media and reliability:** close MoviePy readers after failed audio extraction, discard empty Whisper segments, and ignore outcomes from older circuit-breaker recovery cycles (#427, #438, #425).\n- **Markdown-to-PDF:** honor configured CSS files, preserve relative images, and import successfully when optional WeasyPrint system libraries are missi","is_prerelease":false},"readme":true,"readme_zh":true,"news":[],"stale":false,"snapshot_at":"2026-11-17T21:58:29Z","history_start":"2026-09-06"}