TheDevCharlie/yakob-desktop-assistant

Yakob (ያዕቆብ) - Multilingual Desktop Voice Assistant with Amharic and English Speech Recognition, Neural Male TTS, Alexa-style utilities, and Windows automation # Yakob (ያዕቆብ) ### Intelligent Multilingual Desktop Voice Assistant *Native Amharic (አማርኛ) and English Natural Language Processing, Neural Speech Synthesis, and Windows System Automation.* **Features** • **Architecture** • **Quick Start** • **Voice Commands** • **Voice Training** --- ## Overview **Yakob (ያዕቆብ)** is an enterprise-grade, privacy-conscious desktop voice assistant engineered specifically for bidirectional **Amharic (አማርኛ)** and **English** voice interactions. Combining **real-time Voice Activity Detection (VAD)**, **high-definition neural speech synthesis**, **multi-provider LLM intelligence (Google Gemini 2.5 Flash / Groq Llama 3.3)**, and **direct Windows API automation**, Yakob provides an intuitive, hands-free computing experience packaged inside a modern minimalist dark-themed interface and a sleek floating desktop widget. --- ## System Architecture ```mermaid graph TD User([User Voice Input]) -->|Audio Stream| VAD[Audio Recorder & VAD Engine] VAD -->|WAV Buffer| STT[Speech-to-Text Engine\nam-ET / en-US] STT -->|Transcribed Text| Router{Intent & NLP Router} Router -->|System & App Commands| SysCtrl[System Controller\nApps, Vol, Timer, Screen] Router -->|Alexa Utilities & Trivia| Trivia[Curated Trivia Engine] Router -->|Open-Ended Queries| LLM[LLM Brain\nGemini 2.5 Flash / Groq / OpenAI] SysCtrl -->|Action Result| TTS[Neural TTS Engine\nEdge-TTS HD / ElevenLabs] Trivia -->|Knowledge Result| TTS LLM -->|Conversational Output| TTS TTS -->|Barge-In Capable Audio| Speaker([Speaker Output]) TTS -.->|Interrupt Signal| VAD ``` --- ## Features - **Bidirectional Bilingual NLU**: - Full native recognition of Ge'ez script and Amharic phonology (`am-ET`) alongside American English (`en-US`). - Automatic language identification with hybrid fallback resolution. - **Voice Barge-In / Instant Interruption**: - Full-duplex audio pipeline allows users to interrupt ongoing voice responses at any instant simply by speaking into the microphone. - …