VideoRAG is a desktop app (Vimo) that enables chat with videos using VideoRAG-based retrieval-augmented generation, supporting long videos and multi-modal analysis. It provides a Python backend and Electron frontend workflow, with a focus on long-context video understanding and multi-video analysis.
Collecting history — the radar snapshots this repo daily. The trend line appears after 3 days of data (1 so far).
What it is
VideoRAG: Chat with Your Videos (Vimo Desktop) is a desktop application that enables conversational interaction with videos using VideoRAG technology. It targets long videos and multi-modal content analysis, leveraging a graph-driven knowledge indexing approach and hierarchical context encoding for long sequences.
How it works
VideoRAG combines:
- Graph-Driven Knowledge Indexing: multi-modal knowledge graphs for structured video understanding
- Hierarchical Context Encoding: preserves spatiotemporal visual patterns across long sequences
- Adaptive Retrieval: dynamic retrieval optimized for video content
- Cross-Video Understanding: semantic relationships across multiple videos
Technical highlights include efficient processing on a single RTX 3090 (24GB), structured indexing of long videos into concise knowledge representations, multi-modal retrieval aligning textual queries with visual and audio content, and evaluation on a longer-video benchmark with 134+ hours across various domains.
Getting started
Quick Start shows two options: download Vimo app (Coming Soon for macOS) and run from source. For setup, see Vimo-desktop for complete installation and configuration steps. Quick Overview steps:
- Set up the Python backend environment and start the VideoRAG server
- Launch the Electron frontend application
- Start chatting with your videos!
Recent releases
RELEASES (latest 0): - none
Traction
Stars: 3234; Forks: 461; Open issues: 20
Caveats
License: none listed Created: 2025-02-03; Last push: 2026-03-18