voxai 0.1.0

Voice-driven AI assistant for real-time transcription and Gemini integration.

VoxAI

Voice-driven AI assistant that captures audio from any app (Teams, Zoom, browsers) and streams live, textual answers—ideal for interviews, meetings, and knowledge work.

---

🚀 Quickstart

1. Clone the repo

   git clone https://github.com/rtiwariops/VoxAI.git
   cd VoxAI
   

2. Create & activate a Python virtual environment

   python3 -m venv .venv
   source .venv/bin/activate      # macOS/Linux
   # .venv\Scripts\activate       # Windows PowerShell
   

3. Install Python dependencies

   pip install --upgrade pip
   pip install .
   

4. Install Electron (if not already)

   npm install -g electron
   

5. Configure environment variables

   cp .env.example .env
   

6. Edit .env and provide:

   GENAI_API_KEY=sk-...
   GENAI_MODEL=gemini-1.5-flash
   

7. Run VoxAI

   VoxAI
   

---

🎧 How It Works

• Start Recording
  Click Start to capture audio from your input device (e.g., BlackHole).

• Stop & Transcribe
  Click Stop. Whisper transcribes and shows the full audio clip under Transcript.

• Ask AI
  Click Ask AI to send the transcript to Gemini. Answers stream live in the Answer panel.

• Copy & Share
  Output is plain text—reuse it in RAG/finetune workflows or anywhere else.

---

⚙️ Features

• Universal Audio Capture (BlackHole, etc.)
• Manual Control for long-form Q&A
• One-Shot, Full-Clip Transcription
• Live Token-by-Token AI Streaming
• Configurable Gemini Model (.env)
• Electron-Based Desktop UI

---

📦 Installation Options

From source:

git clone https://github.com/YourUsername/VoxAI.git
cd VoxAI
pip install -e .
VoxAI

From PyPI:

pip install VoxAI
VoxAI

---

🛠 Configuration

Set in .env:

Variable	Description	Example
GENAI_API_KEY	Google Generative AI (Gemini) API key	sk-…
GENAI_MODEL	Gemini model to use	gemini-1.5-flash

See .env.example for reference.