skills.rest
Trac-Systems
Trac-SystemsOFFICIAL
Explore all skills by this author →
WebsiteGitHub

local-voice-gateway

Run a local ASR-to-LLM-to-TTS voice pipeline on macOS and Linux.

3|1|Updated Feb 22, 2026
Search Tags:#llm#macos#linux#local#tts#gateway#asr
One-click install
npx skills add https://github.com/Trac-Systems/clawfinger-gateway --skill local-voice-gateway
Or copy as Structured Prompt for Agent▼
Please help me install this Agent Skill.
Skill: local-voice-gateway
Source: https://github.com/Trac-Systems/clawfinger-gateway/tree/main/skills/voice-gateway
Command: npx skills add https://github.com/Trac-Systems/clawfinger-gateway --skill local-voice-gateway

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve?

Local voice gateway for Clawfinger provides a self-contained, on-device gateway that runs the entire ASR → LLM → TTS pipeline without relying on cloud services, enabling offline development and testing.

Core Features & Use Cases

  • Local gateway (FastAPI) that exposes REST API endpoints, a control center UI, and an agent WebSocket interface for integration with the Clawfinger Android app.
  • On-device inference stack (ASR via mlx_audio Parakeet, LLM via mlx-lm or remote endpoints, TTS via Kokoro for English and Piper for German).
  • Cross-platform support (macOS and Linux) with ADB reverse port forwarding for phone connectivity.
  • Comprehensive installation, configuration, operation, and troubleshooting guidance, including security and network-local hosting constraints.

Quick Start

Create a Python virtual environment, install the required dependencies, copy the example config to config.json, and start the gateway with the provided start script.

Frequently Asked Questions about local-voice-gateway

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I run a local ASR LLM TTS pipeline offline on macOS or Linux?▼

A local voice gateway runs the full ASR LLM TTS pipeline entirely on-device for macOS and Linux. It uses a FastAPI server with REST endpoints and a WebSocket interface to handle speech recognition, text generation, and audio synthesis offline.

Can I use a local TTS gateway with an Android app over ADB?▼

Yes, the local voice gateway supports Android app integration through ADB reverse port forwarding. This connects the Clawfinger app to your local WebSocket agent interface, allowing on-device voice processing without cloud services.

What models are used for local speech recognition and text to speech?▼

Local speech recognition uses the mlx_audio Parakeet model, while text to speech uses Kokoro for English and Piper for German. The LLM component runs via mlx-lm for local inference or through remote endpoints if configured.

Do I need Python 3.12 to set up a local voice gateway?▼

Yes, Python 3.12 or higher is required to set up the local voice gateway. You also need macOS or Linux, ADB for phone connectivity, and approximately 4 GB of storage for the required model data.

How do I configure and start a local FastAPI voice gateway?▼

To start the local FastAPI voice gateway, create a Python virtual environment, install dependencies, and copy the example configuration to config.json. Then run the provided start script to launch the server and control center UI.

Related Skills

Explore More Software Engineering Skills→
crazymsncrazymsn
22

get-available-resources

Detect CPU, GPU, memory, and disk resources and write a .claude_resources.json report.

Community
Intermediate
skydovesskydoves
303

connecting-over-wifi

Pairs Android 11+ devices over Wi-Fi and connects to the post-pairing debug port via mDNS.

Community
Advanced
yankeeinlondonyankeeinlondon
3

tts

Integrate neural TTS models across Rust, Python, and TypeScript.

Community
Advanced
HuuBarHuuBar

asr-transcribe-to-text

Transcribe audio and video files to text using Qwen3-ASR.

Community
Advanced