StreamRAG enables developers to build GPT-powered agents for searching and retrieving information from video libraries. It allows users to upload multiple videos, query them for summarized text answers or key insights, and publish these searchable collections on the ChatGPT store.

How It Works

StreamRAG leverages VideoDB for video indexing and retrieval. Users upload video files or links, which are processed and stored by VideoDB. A Flask server acts as an API endpoint, allowing a GPT agent to query this video library. The system supports Retrieval-Augmented Generation (RAG) to provide summarized text responses based on video content.

Quick Start & Requirements

• Install dependencies: pip install -r requirements.txt
• Obtain a VideoDB API key (free tier available).
• Set VIDEO_DB_API_KEY environment variable.
• Run the Flask server: python app.py
• For ChatGPT store integration, deploy the Flask server and update openapi.yaml with the server URL.
• Official documentation and demo videos are linked in the README.

Highlighted Details

• Enables creation of searchable video libraries.
• Supports real-time video responses and compilations.
• Facilitates publishing on the ChatGPT store.
• Provides summarized text answers via RAG.

Maintenance & Community

The project welcomes contributions through issues, brainstorming, and code improvements. Standard GitHub pull request workflow is outlined.

Licensing & Compatibility

The repository's license is not explicitly stated in the provided README. Compatibility for commercial use or closed-source linking is not detailed.

Limitations & Caveats

The project is currently focused on a Flask backend and lacks integrations with common deployment pipelines like Heroku or Replit. Support for cloud storage providers like Dropbox or Google Drive is also planned for future integration.