A
licenseNot graded
qualityC
maintenanceEnables AI agents to speak and listen in real-time with interruption handling, using local ML models and hot-swappable adapters.
32 npm
MIT
genpark-voice-turn-taking-endpointing-detector-skill
by Alpha-Park
Click on "Deploy Server".
Wait a few minutes for the server to deploy. Once ready, it will show a "Started" state.
In the chat, type
@followed by the MCP server name and your instructions, e.g., "@genpark-voice-turn-taking-endpointing-detector-skillEnable barge-in detection and tune endpointing sensitivity"
That's it! The server will respond to your query, and you can continue using it as needed.
Here is a step-by-step guide with screenshots.
GenPark AI Agent Skill - Voice Turn-Taking & Endpointing Detector
Acoustic voice activity detection (VAD), dynamic silence endpointing, and barge-in interruption arbitrator for real-time conversational voice agents.
Verified by GenPark AI and compatible with Model Context Protocol (MCP).
Architecture Diagram
graph TD
A[Real-time PCM Audio Stream] --> B[Energy & VAD Frame Analyzer]
B --> C{Agent Currently Speaking?}
C -->|Yes & Speech Energy > Threshold| D[Barge-In Interrupt Triggered: Cut Agent Audio]
C -->|No| E[Accumulate Speech & Silence Frames]
E --> F{Silence Duration >= Endpointing Threshold?}
F -->|Yes| G[Decision: Turn Completed -> Dispatch LLM Synthesis]
F -->|No| H[Decision: Continue Listening]Related MCP server: genpark-realtime-voice-turn-taking-orchestrator-skill
Features
Low-Latency VAD Heuristics: Operates on sub-frame time slices for immediate interruption handling.
Dynamic Endpointing: Automatically balances conversational fluidity against premature turn cutoff.
Zero External Dependencies: Pure Python standard library implementation.
This server cannot be deployed
Maintenance
ActivityMaintained
ResponsivenessNo issues
Related MCP Connectors
Hosted speech-to-text + speech emotion/tone analysis for agents. No install; trial keys built in.
Transcribe audio & video to text for AI agents: 100+ languages, speaker labels, webhooks.
Audio for your agent: transcribe, speak, translate, summarise, plus sound effects and music.
Speech, transcription, voice agents, Trace, Recap, dubbing and narration with browser OAuth.
Related MCP Servers
- FlicenseNot gradedqualityBmaintenanceEnables low-latency conversational turn-taking and barge-in handling for real-time voice AI interactions, orchestrating speech flow with sub-100ms responsiveness.8-
- FlicenseNot gradedqualityBmaintenanceEnables real-time detection of speech turn completion, distinguishing mid-sentence hesitation pauses from finished turns for voice agents and conversational audio pipelines.7-
- FlicenseNot gradedqualityBmaintenanceEnables real-time voice agents and streaming audio pipelines to detect conversational turn completion, distinguishing finished speech turns from mid-sentence hesitation pauses using acoustic VAD and semantic scoring. It provides low-latency decision and telemetry output through a native MCP interface for integration with MCP-compliant clients.7-
