AhmadHassan-BTed/B
AUTONOMOUS AI DESKTOP COMPANION
An autonomous AI assistant living on your desktop. Observes the screen, tracks the active window, and proactively interacts via voice and an animated interface.
// KEY FEATURES
- Proactive behavior: initiates dialogue based on screen context
- 60fps event bus: perception, cognition, expression in real time
- Local ONNX TTS + OCR: voice and vision without cloud APIs
- Reactive UI with facial expressions: floating overlay with emotional states
// DEEP DIVE: HOW IT WORKS
Architecture built on async EventBus: sensors (WindowTracker, SemanticSensor, VisionSensor) pass context to CognitiveEngine, LLM generates response, VoiceEngine (ONNX TTS + DSP vocoder), FaceRenderer, and KinematicsEngine voice and animate avatar. 60 FPS loop reacts to window switches and screen content.
Proactive: doesn't wait for commands, analyzes active window semantics via UIA tree, falls back to OCR. State machine drives emotions and curiosity, AutonomyEngine triggers thoughts. Modules fully isolated, communicate only via EventBus — easy to test and swap components.
Requires Windows (UIA for semantic analysis), Python, and local LLM. Voice runs on ONNX offline, vision via OCR. No install instructions in README; early project, may have dependency and performance issues on weak hardware.
// USE CASES
Best for: For AI agent developers and local LLM enthusiasts working with autonomous desktop assistants
- Watch code in IDE and get contextual hints without switching windows
- Receive voice comments on documentation or errors while working
- Explore autonomous agent behavior with emotional state on your desk
New repositories every 30 minutes
REDDYX AI scans GitHub 24/7 and ships the best AI/ML/Web3 projects to Telegram.
Join on Telegram