Voco V2

The Voice-Native Coding Orchestrator — Sub-300ms Voice-to-Action

Voco is a local-first, zero-trust voice interface for developers. Speak naturally and Voco will search your codebase, generate code, run terminal commands, and explain results — all with human-in-the-loop approval.

Quick Start (5 Minutes)

# 1. Clone and install
git clone https://github.com/Radix-Obsidian/Voco-ai.git
cd Voco-ai/services/mcp-gateway
npm install

# 2. Set up environment
cp .env.example .env
# Add your API keys from https://voco.ai/dashboard

# 3. Start development
npm run dev

Features & Capabilities

🎙️ Voice-to-Context Engine

• Sub-300ms voice transcription via Deepgram
• Barge-in support with Silero VAD
• Instant voice feedback via Cartesia TTS
• Context-aware LangGraph state machine

🔒 Zero-Trust MCP Gateway

• Tauri v2 secure sandbox
• Human-in-the-loop terminal approval
• Filesystem scope validation
• Row-level security via Supabase

💡 Enterprise Features

• "Seat + Meter" billing ($19/mo + $0.02/turn)
• Team workspaces & shared context
• Audit logs & usage analytics
• SOC 2 compliance ready

Architecture

Voco V2 is a monorepo with two independent runtimes:

services/
├── mcp-gateway/        # Local frontend & MCP execution sandbox
│   ├── Tauri v2 (Rust) # Zero-trust security layer
│   ├── React 18       # Modern UI with Shadcn components
│   └── TypeScript     # Type-safe codebase
│
└── cognitive-engine/   # Cloud reasoning & audio engine
    ├── LangGraph      # Stateful AI orchestration
    ├── Silero-VAD     # Voice activity detection
    ├── FastAPI        # WebSocket bridge
    └── Python 3.12+   # Async runtime

MCP Gateway (services/mcp-gateway/)

The local desktop application built with Tauri. Provides:

• Zero-trust MCP execution sandbox
• Human-in-the-loop terminal approval
• Filesystem scope validation
• WebSocket bridge to cognitive engine
• Supabase auth & RLS integration
• Modern React UI with Shadcn/UI

Stack: Tauri v2 • React 18 • TypeScript • Vite • Tailwind CSS 3 • Shadcn/UI • Supabase • Framer Motion

Cognitive Engine (services/cognitive-engine/)

The cloud reasoning service that handles:

• Voice transcription (Deepgram)
• Text-to-speech (Cartesia)
• LangGraph state machine
• Tool execution & validation
• Background job queue
• Usage metering & billing

Stack: Python 3.12+ • FastAPI • LangGraph • Silero-VAD • Supabase • Stripe

Getting Started

Prerequisites

• Node.js >= 18
• npm (or bun)

Setup

# Clone the repository
git clone https://github.com/Radix-Obsidian/Voco-ai.git
cd Voco-ai

# Install MCP Gateway dependencies
cd services/mcp-gateway
npm install

# Configure environment
cp .env.example .env
# Edit .env with your Supabase credentials

# Start the dev server
npm run dev

The app will be available at http://localhost:8080.

Environment Variables

Variable	Description
VITE_SUPABASE_URL	Your Supabase project URL
VITE_SUPABASE_PUBLISHABLE_KEY	Your Supabase anon/public key

Project Structure

.
├── Docs/                          # Product & technical documentation
│   ├── PRD.md                     # Product Requirements Document
│   ├── TDD.md                     # Technical Design Document
│   ├── SDD.md                     # System Design Document
│   ├── Core-Features-List.md      # Feature inventory
│   └── GTM.md                     # Go-to-Market strategy
├── services/
│   └── mcp-gateway/
│       ├── src/
│       │   ├── assets/            # Branding assets (logo, icon)
│       │   ├── components/
│       │   │   ├── ui/            # Shadcn UI primitives (48 components)
│       │   │   ├── AuthModal.tsx   # Supabase email/password + Google OAuth
│       │   │   ├── Header.tsx      # App header with logo
│       │   │   └── ProtectedRoute.tsx
│       │   ├── hooks/             # React hooks (auth, settings, projects, toast)
│       │   ├── integrations/
│       │   │   └── supabase/      # Supabase client & generated types
│       │   ├── lib/               # Utilities (cn helper)
│       │   └── pages/             # Landing, AppPage, NotFound
│       ├── tailwind.config.ts
│       ├── vite.config.ts
│       └── components.json        # Shadcn configuration
├── .windsurfrules                 # AI agent architectural guardrails
└── README.md

Documentation

Detailed documentation lives in the Docs/ directory:

• PRD.md — Product requirements and feature parity with V1
• TDD.md — LangGraph & streaming audio architecture
• SDD.md — Zero-trust MCP & human-in-the-loop sandbox design
• Core-Features-List.md — Complete feature inventory
• GTM.md — Go-to-market strategy

Community

• Discord: Join the community
• GitHub Issues: Report bugs or request features
• Contributing: See CONTRIBUTING.md for guidelines
• Code of Conduct: See CODE_OF_CONDUCT.md
• Security: See SECURITY.md for responsible disclosure

Contributing

We welcome contributions! Please read our Contributing Guide before submitting a pull request.

1. Fork the repository
2. Create a feature branch from main
3. Make your changes and add tests
4. Submit a pull request with a clear description

License

Licensed under the Apache License 2.0.