voice-ai

local-voice-ai: Real‑time Voice Interaction for Your Startup Apps

Name: local-voice-ai: Real‑time Voice Interaction for Your Startup Apps
Author: ShayneP

Leverage local-voice-ai to add on‑device speech recognition and AI responses using llama‑cpp, Kokoro, Nemotron STT, and LiveKit. Ideal for founders building privacy‑first voice features.

From the repo · “Local voice AI powered by llama-cpp, Kokoro, Nemotron STT, and LiveKit.”

725 stars217 forksTypeScriptGuide quality 9/10 · reviewed Oct 22Updated 8/31/2026100% free · open source

What it does

Local voice AI powered by llama-cpp, Kokoro, Nemotron STT, and LiveKit allows founders to integrate voice AI into their applications without relying on cloud services.

When to use it

•When developing voice-controlled applications for areas with poor internet connectivity
•When integrating voice AI into an existing application with strict data privacy requirements
•When prototyping a voice-controlled product that requires low-latency responses

Ready-to-paste prompt

Use the `local-voice-ai --recognize 'Hello, how are you?'` command to test speech recognition capabilities

Heads up: Make sure to install the required dependencies, including `libllvm-dev` and `libopus-dev`, to successfully compile the llama-cpp library and avoid compilation errors

Saves to your device

Use with Claude

New

Skip the builder — one click puts this in Claude, Cursor, Antigravity and more.

✅ Light setup

Installs with a command or two; your AI agent can do it for you.

Try it instantly — no install

Claude Code

mkdir -p ~/.claude/skills/local-voice-ai && curl -fsSL https://workflowstacks.com/api/skills/local-voice-ai/claude-skill -o ~/.claude/skills/local-voice-ai/SKILL.md

Open in another AI app

Opens the app with this repo with the prompt ready to go — no copy-paste needed.

Connect the whole catalog (MCP)

claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

How local-voice-ai: Real‑time Voice Interaction for Your Startup Apps works

Codeflow

Free to inspect

Local-voice-ai: Real‑time Voice Interaction for Your Startup Apps is a medium Python project (~6.8k lines across 62 code files, plus 18 test files). Setup is light: runs with Docker. Last commit 2 months ago, MIT license, has a test suite.

Size

Medium codebase

~6.8k lines · 62 code files · 57 min skim

Setup

Light setup

Runs with Docker.

Runs on

Python · Docker

No API keys detected

Python 77%TypeScript 20%Dockerfile 1%CSS 1%

Where to start reading

1
README.md
Start here — what it does and how to install it
2
run.py
Where the program starts running
3
pyproject.toml
Dependencies and the commands it exposes
4
local_voice_ai/__init__.py
Inside local_voice_ai/ — the main logic begins here

What's in each folder

local_voice_ai/Core code — the actual logic25 files

frontend/Frontend / UI68 files

tests/Tests — proof it works18 files

.github/CI / automation (GitHub Actions)1 files

READMEHas testsThin docsCI checksDocker readyMIT licenseUpdated 2 mo ago

Use with Claude

New

Skip the builder — one click puts this in Claude, Cursor, Antigravity and more.

✅ Light setup

Installs with a command or two; your AI agent can do it for you.

Try it instantly — no install

Claude Code

mkdir -p ~/.claude/skills/local-voice-ai && curl -fsSL https://workflowstacks.com/api/skills/local-voice-ai/claude-skill -o ~/.claude/skills/local-voice-ai/SKILL.md

Open in another AI app

Opens the app with this repo with the prompt ready to go — no copy-paste needed.

Connect the whole catalog (MCP)

claude mcp add --transport http workflowstacks https://workflowstacks.com/api/mcp

Adds a WorkflowStacks connector to Claude Code: search and load any skill here by chatting.

Quick Actions

Details

Creator

ShayneP

Language

TypeScript