Back to search
Top Contributors
254
commits
Your AI assistant that never forgets and runs 100% privately on your computer. Leave it on 24/7 - it learns your preferences, helps with code, manages your health goals, searches the web, and connects to any tool via MCP servers (e.g. home automation). No subscriptions, no cloud, just say "Jarvis" anywhere in your sentence and talk naturally.
Stars
136
Forks
31
Watchers
136
Open Issues
32
Overall repository health assessment
No package.json found
This might not be a Node.js project
254
commits
fix: add SVG indicator icons for checkboxes, radio buttons, and spinners
c16a0dcView on GitHubfeat: add auto-generated settings UI for config editing (#67)
0ba2c55View on GitHubfix: fall back to device native sample rate when 16 kHz is rejected (Linux) (#120)
d0941ceView on GitHubfeat: expose gemma3n:e2b and gemma3n:e4b as explicit model choices in setup wizard (#107)
4e1fcd4View on GitHubfix: preserve lock file PID so 'Close Existing & Start New' works (#108)
ef97e40View on GitHubrefactor: move mocked greeting/tool tests from evals to unit tests (#109)
bd8c215View on GitHubrefactor: decouple English constraint from model size, tie to TTS engine only
b7ba31fView on GitHubfeat: default Whisper model to medium multilingual and fix model specs
7b7a54dView on GitHubfix: use os.add_dll_directory for NVIDIA DLLs and improve fallback message
e297e70View on GitHubfeat: automatic NVIDIA CUDA GPU acceleration for Whisper on Windows
dfbffe7View on GitHubfix: CUDA detection checks for both cuBLAS and cuDNN on Windows
5777112View on GitHubfeat: switch default small model from llama3.2 to gemma3n with automatic tool-calling fallback
d11bc57View on GitHubfeat: Windows installer via Inno Setup with onedir build (#99)
5d7dbcdView on GitHub