Getting Started

MCP Voice Engine is a deterministic, streaming-first prosody engine for expressive voice synthesis.

Prerequisites

• Node.js 20 or later
• npm

Install and build

git clone https://github.com/mcp-tool-shop-org/mcp-voice-engine.git
cd mcp-voice-engine
npm i
npm run build

Run tests

The full test suite covers determinism, meaning contracts, and real-time performance:

# Full suite
npm test

# Specific suites
npm run test:meaning       # communicative behavior guardrails
npm run test:determinism   # hash-based regression tests
npm run bench:rtf          # real-time factor benchmark
npm run smoke              # end-to-end smoke test

Repository structure

The monorepo contains one primary package today, with clean separation for future synthesis backends:

mcp-voice-engine/
└── packages/
    └── voice-engine-dsp/    # Core DSP + streaming prosody engine
        ├── docs/            # Architecture, meaning contract, debugging
        ├── src/             # Source code
        └── tests/           # Meaning, determinism, and benchmark suites

Key documentation

The primary docs live alongside the DSP package:

• packages/voice-engine-dsp/docs/STREAMING_ARCHITECTURE.md — causal processing model
• packages/voice-engine-dsp/docs/MEANING_CONTRACT.md — prosody behavior specification
• packages/voice-engine-dsp/docs/DEBUGGING.md — debugging guide
• Reference_Handbook.md — full API and concepts reference