GH Repository · HKUDS
VideoRAG
[KDD'2026] "VideoRAG: Chat with Your Videos"
- stars
- 3,367
- 30-day movement
- +83/day
- Related entries
- 60
- Connections
- 1
Pythonretrieval-augmented-generationlong-video-understandingmulti-modal-llmsllmsraglarge-language-models
VideoRAG is a Python research project from HKUDS, published at KDD'2026, that implements retrieval-augmented generation over video content under the tagline "Chat with Your Videos". Its tags indicate it targets long-video understanding and multi-modal LLMs, combining large language models with RAG techniques for video.
Use it when you want to apply retrieval-augmented generation so you can query and chat with video content using LLMs.
Use it to
- Chat with video content using an LLM
- Apply RAG techniques to long videos
- Explore multi-modal LLM retrieval pipelines
- Reproduce the KDD'2026 VideoRAG research
For Researchers and developers working on multimodal RAG systems
- Role
- rag
- Language
- Python
- Licence
- custom (licence file present)
- Forks
- 476
- Open issues
- 20
- Last push
- 2026-03-18
topicsragvideo-understandinglarge-language-modelsmulti-modalretrieval-augmented-generation