# HKUDS/VideoRAG

[KDD'2026] "VideoRAG: Chat with Your Videos"

Repository: https://github.com/HKUDS/VideoRAG
Canonical: https://ross.abutalabs.com/products/videorag
Homepage: https://arxiv.org/abs/2502.01549
Language: Python
License: NOASSERTION
License Family: other
Topics: large-language-models, llms, long-video-understanding, multi-modal-llms, rag, retrieval-augmented-generation
Last push: 2026-03-18T08:33:07+00:00

## Health v2 (maintenance only)
Score: 53/100 (v2, computed 2026-09-02T17:46:02.011165+00:00)
- activity 72, release rhythm 35, longevity 41
- inputs: {"age_days": 576, "days_push": 168, "days_rel": null, "gap_med": null, "n_releases_24m": 0}
- flags: no_releases, no_license
- formula: round(0.45*activity + 0.35*rhythm + 0.20*longevity); archived -> min(score, 10)

## Adoption (not part of the score)
Stars 3337, forks 478 (observed 2026-08-28T04:07:56.791417+00:00)

## What it is
VideoRAG is a retrieval-augmented generation framework for chatting with and understanding extremely long-context videos, using a dual-channel architecture combining graph-based textual knowledge grounding with multi-modal visual context encoding. It includes a LongerVideos benchmark and a Vimo desktop application.

## Use cases
- chat with my videos using an llm
- build a rag system over long video collections
- search and query across many hours of video content
- understand extremely long videos with multi-modal retrieval
- build a knowledge graph spanning multiple videos
- benchmark long video understanding methods

## When to choose
- you need question answering over very long or many videos
- you want graph-based retrieval combined with visual multi-modal encoding
- you need a research baseline for long-video RAG

## When to avoid
- you only need text-only RAG over documents
- you need a lightweight production video search without LLM dependencies
- you require a permissively licensed library (license is non-standard)

## Facets
- artifact type: framework
- maturity: active
- function: rag, machine-learning, video-processing, llm-inference, search-engine, chatbot
- domain: large-language-models, artificial-intelligence, machine-learning
- platform: python, cross-platform
- tags: multi-modal, long-video-understanding, knowledge-graph, video-qa, desktop-app, research, retrieval-augmented-generation, video, gpu

## Member repositories
- HKUDS/VideoRAG (main) score 53

## Provenance
- Observed fields: from GitHub, fetched 2026-08-28T04:07:56.791417+00:00.
- Health v2: computed from the inputs above; adoption is never an input.
- Inferred fields (summary, facets, guidance): AI-extracted, prompt v1, taxonomy v1, on 2026-08-29T18:41:06.539173+00:00, confidence not recorded.
  - readme: https://github.com/HKUDS/VideoRAG (fetched 2026-08-28T04:07:56.791417+00:00, sha a7b7c75bb352)
  - homepage: https://arxiv.org/abs/2502.01549 (fetched 2026-08-29T09:34:29.681621+00:00, sha 1ebc9cafc122)
  - site_page: https://info.arxiv.org/about/donate.html (fetched 2026-08-29T09:34:29.690802+00:00, sha cca9c3a11c56)
  - site_page: https://info.arxiv.org/about/ourmembers.html (fetched 2026-08-29T09:34:29.694040+00:00, sha 47cbc55ff1de)
  - site_page: https://info.arxiv.org/about (fetched 2026-08-29T09:34:29.696258+00:00, sha a1f16f915a9a)
  - site_page: https://info.arxiv.org/labs/index.html (fetched 2026-08-29T09:34:29.692447+00:00, sha b14a8d05a0ec)
- Data as of 2026-08-30T08:39:29.467469+00:00.