RayP11/Friday---Llama-Offline-AI-Assistant
📦 Open Source ProjectRayP11
A fully local, privacy-focused AI voice assistant powered by Ollama, LangChain, and Whisper.
Friday represents a significant step toward private, local-first AI interaction. Built with Python, the project integrates several powerful open-source technologies to create a seamless voice-to-voice loop. At its core, it uses Ollama to run large language models locally, allowing for sophisticated reasoning without external connectivity. The speech recognition pipeline is handled by Whisper, providing high-accuracy transcription, while Sapi5 provides the voice output. The architecture is modular, utilizing LangChain to manage the conversation flow and context. This setup is ideal for developers or privacy enthusiasts who want to build or deploy a voice assistant that keeps all data on the local hardware, mitigating the risks associated with cloud-based voice processing services. The project is highly extensible, allowing users to swap models or integrate additional local tools via the LangChain framework.
💡Highlights
- ├─100% local processing, no cloud API
- ├─Integrates Ollama, Whisper, and LangChain
- └─Sapi5-based offline voice synthesis
🎯For
- ├─Privacy-focused developers
- └─AI hobbyists