IliasHad/edit-mind

Local-first Video Knowledge Base. Index your video library with multi-modal analysis (YOLO, DeepFace, Whisper), search semantically via natural language, Docker-ready.

What it solves

Edit Mind creates a local, AI-powered knowledge base for your video library. It eliminates the need to manually scrub through hours of footage to find specific moments by allowing you to search for spoken words, objects, faces, or specific scenes using natural language.

How it works

The system runs as a background service that monitors your video folders. It uses a combination of AI models to analyze video content:

  • Transcription: Converts speech to text using OpenAI Whisper.
  • Visual Analysis: Performs face recognition, object detection, and text detection.
  • Semantic Search: Uses multi-model embeddings and ChromaDB (a vector database) to enable natural language queries across the indexed metadata.
  • LLM Integration: Supports local models via Ollama or cloud models via Google Gemini for NLP tasks.

Who it’s for

Video editors and content creators who need a "second brain" to organize and retrieve specific clips from large volumes of local video archives while maintaining privacy.

Highlights

  • Fully Local: Runs on your own hardware via Docker, ensuring your videos stay private.
  • Cuda Support: Includes specific configurations for NVIDIA GPUs to accelerate processing.
  • Multi-Model Indexing: Combines transcription, frame analysis, and embeddings for comprehensive search.
  • Broad Compatibility: Works on any computer or server with Docker installed.