Hi All,
Please find this week’s development progress and planning call livestream recording, followed by an AI powered summary of the discussion below.
Questions, comments, concerns, and ideas welcome as always.
Just hit the reply button.
Thanks,
Dave W.
Recording:
AI Powered Summary:
Development Updates: RAG (Retrieval-Augmented Generation)
- Mac Porting: The team successfully ported the RAG system to Mac. The install process is notably efficient, completing in approximately 2 to 3 minutes on a Mac Mini.
- Cross-Platform Compatibility: With the Mac port complete, the foundation for RAG is now functional across Linux, Windows, and Mac.
- Local Processing: All document parsing, chunking, and vector database storage occur locally. This allows you to chat with your private data—like tax info or company documents—entirely offline and securely.
- Freedom of Choice: Users are not locked into specific models; you can choose your own embedding models (like Nomic) and LLM providers (like Ollama or Open Router) to customize how the system handles your data.
The Week Ahead & Future Roadmap
- UI/UX Polishing: The upcoming week will focus on “polish,” specifically adding visual indicators to show when services are starting or stopping so users aren’t left guessing.
- Management Features: New controls are being added to allow users to easily remove document collections and to provide “Are you sure?” warnings when making significant changes like switching embedding models.
- Logging & Debugging: A new system information tool is being developed. This will allow “owners” to generate technical logs to help the team troubleshoot bugs without ever exposing personal data or private chat history.
- Performance Testing: As the foundation is now stable, the team will begin stress-testing the system with larger document sets (50+ files) to identify and eliminate latency pain points.
- Installers: Work continues on simplifying the initial setup, including the development of a Docker version and refining the beta Windows and Mac installers.
Get Involved:
- Documentation: Visit docs.braindrive.ai for the Owner’s Manual and building guides.