HKUDS/VideoRAG
[KDD'2026] "VideoRAG: Chat with Your Videos"
Python · ★ 2,108 · Fork 293 · 在 GitHub 查看
在手机上浏览源码、订阅每日 AI 趋势日报与项目架构深度解读:获取 CopoHub 客户端(iOS & HarmonyOS)→
GitHub Trending 上榜记录
HKUDS/VideoRAG 已在 GitHub Trending 每日榜单上榜 1 天,最高排名第 15,首次上榜 2026-01-09,最近一次上榜 2026-01-09。
| 日期 | 排名 | Star | 当日增长 |
|---|---|---|---|
| 2026-01-09 | #15 | 2,108 | +127 |
同期上榜的项目
- ChromeDevTools/chrome-devtools-mcp – Chrome DevTools for coding agents
- anthropics/claude-code – Claude Code is an agentic coding tool that lives in your terminal, understands your codebase, and helps you code faster by executing routine tasks, explaining complex code, and handling git workflows - all through natural language commands.
- nothings/stb – stb single-file public domain libraries for C/C++
- MiroMindAI/MiroThinker – MiroThinker is an open source deep research agent optimized for research and prediction. It achieves a 80.8% Avg@8 score on the challenging GAIA benchmark.
- protocolbuffers/protobuf – Protocol Buffers - Google's data interchange format
- thedotmack/claude-mem – Persistent Context Across Sessions for Every Agent – Captures everything your agent does during sessions, compresses it with AI, and injects relevant context back into future sessions. Works with Claude Code, OpenClaw, Codex, Gemini, Hermes, Copilot, OpenCode + More
HKUDS/VideoRAG 项目深度剖析与商业评估报告
HKUDS/VideoRAG 项目深度剖析与商业评估报告
报告日期:2026-08-22 | 分析对象:VideoRAG (Vimo Desktop) 仓库状态:3,328 Stars / 474 Forks / 21 Open Issues | 最后推送 2026-03-18
---
1. 核心定位与商业落地价值
1.1 解决的根本痛点
VideoRAG 瞄准的是长视频多模态理解这一业界公认的硬骨头。传统视频理解方案(如直接对视频逐帧采样送入多模态大模型)存在三个致命缺陷:Token 爆炸(一小时视频可产生数十万 Token)、时序信息丢失(独立帧采样无法建模事件因果链)、检索精度低下(稀疏采样导致关键帧遗漏)。VideoRAG 通过图驱动知识索引 + 层次化上下文编码,将数百小时的视频蒸馏为结构化知识图谱,在推理时仅需检索相关子图,从根本上绕开了"长序列 Transformer 自注意力 O(n²) 复杂度"的物理瓶颈。
1.2 企业级商业落地场景
| 场景 | 价值主张 | 目标客户 | |------|---------|---------| | 视频审查与合规 | 对数百小时监控/会议录像进行语义级检索,替代人工抽检 | 金融、政府、安防 | | 媒体资产管理与剪辑 | 基于自然语言精准定位"某演员穿红裙的片段",大幅降低人工编目成本 | 广电、影视后期、MCN | | 在线教育 | 对整学期课程视频建立知识索引,学生可对话式提问 | K12、高校、企业培训 | | 多模态客服质检 | 分析客服与客户的视频通话记录,自动提取服务态度、话术合规性 | 保险、银行、电商 |
在 README 中展示上榜徽章
[](https://github.fq6825.top/repo/HKUDS/VideoRAG)