Developer Builds Offline Desktop AI Assistant Jansky
Mayukh Bagchi recently built Jansky, an offline desktop AI assistant running on a Raspberry Pi 5 that listens, transcribes, reasons, and speaks entirely locally. It uses Whisper.cpp for speech recognition, Ollama hosting a compact Qwen 2.5 model (1.5 billion parameters), and Piper for neural TTS, with optional cloud fallback. The project demonstrates inexpensive, privacy-preserving on-device assistants are practical today.
Key Points
- 1Builds Jansky offline assistant on Raspberry Pi 5 with touchscreen, mic, speaker, GPIO-powered display
- 2Uses local Whisper.cpp speech recognition and Ollama-hosted Qwen 2.5 (1.5B) to avoid cloud processing
- 3Enables developers to deploy private, always-on voice agents with inexpensive hardware and open software
Scoring Rationale
Practical, reproducible on-device assistant demonstration; limited novelty and single-source hobbyist project constrain broader industry impact.
Sources
Public references used for this report.
Practice interview problems based on real data
1,625 SQL & Python problems across 15 industry datasets — the exact type of data you work with.
Try 250 free problems