Kuon
久远：一个开发中的大模型语音助手，当前关注易用性，简单上手，支持对话选择性记忆和Model Context Protocol (MCP)服务。 KUON:A large language model-based voice assistant under development, currently focused on ease of use and simple onboarding. It supports selective memory in conversations and the
What is Kuon?
Kuon is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to 久远：一个开发中的大模型语音助手，当前关注易用性，简单上手，支持对话选择性记忆和model context protocol (mcp)服务。 kuon:a large language model-based voice assistant under development, currently focused on ease of use and simple onboarding. it ...
久远：一个开发中的大模型语音助手，当前关注易用性，简单上手，支持对话选择性记忆和Model Context Protocol (MCP)服务。 KUON:A large language model-based voice assistant under development, currently focused on ease of use and simple onboarding. It supports selective memory in conversations and the
This server falls under the Communication category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- 久远：一个开发中的大模型语音助手，当前关注易用性，简单上手，支持对话选择性记忆和Model Context Protoc
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx kuonConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use Kuon
Kuon (久远) is a voice AI assistant built in Python that combines speech-to-text, text-to-speech, and LLM dialogue into a single system with optional selective conversation memory and MCP server support. It is designed for ease of use and supports connecting to OpenAI-compatible APIs (including GPT-4, GPT-3.5, and self-hosted models) for its dialogue engine. Kuon's three independent modules — speech recognition (via FunASR), TTS (via online API or offline model), and the LLM dialogue manager — can each be configured or replaced independently, making it a flexible local voice assistant foundation.
Prerequisites
- Python 3.10 with Conda or venv for environment management
- An OpenAI API key or access to an OpenAI-compatible LLM endpoint
- FunASR service running locally for speech recognition (optional if using text-only mode)
- A microphone and speaker/audio output for voice input/output
Clone the repository
Clone the Kuon repository to your local machine.
git clone https://github.com/lissettecarlr/kuon.git
cd kuonCreate a Python 3.10 environment
Create and activate a dedicated Conda environment for Kuon.
conda create -n kuon python=3.10
conda activate kuon
pip install -r requirements.txtConfigure the LLM connection
Edit llm/config.yaml to set your LLM endpoint URL, API key, model type (e.g. gpt-4 or gpt-3.5-turbo), and the character prompt for the assistant's personality.
Configure speech recognition (optional)
If you want voice input, edit kuonasr/config.yaml to point at your FunASR service URL. Skip this step to use text-only input mode by setting audio_input_sw: false in the root config.yaml.
Configure text-to-speech (optional)
Edit kuontts/config.yaml to set your TTS endpoint and speaker settings. Both online API and offline model options are supported.
Run component checks and start
Run the check script first to verify all configured components are working, then start the assistant.
python check.py
python kuon.pyKuon Examples
Root configuration options
The root config.yaml controls which I/O modes are active at startup. Set these flags to enable or disable voice and text input/output.
{
"mcpServers": {
"kuon": {
"command": "python",
"args": ["/path/to/kuon/kuon.py"]
}
}
}Prompts to try
Once Kuon is running, you can speak or type these to interact with the voice assistant.
- "What did we talk about earlier?" (tests selective memory recall)
- "Remember that I have a meeting at 3pm tomorrow."
- "Tell me today's weather in Tokyo."
- "Translate this sentence into Japanese: Hello, how are you?"
- "What is the capital of France?"Troubleshooting Kuon
LLM API connection fails on startup
Verify your API key and endpoint URL in llm/config.yaml. Ensure the model name matches exactly what your provider expects (e.g. 'gpt-4' vs 'gpt-4-turbo'). Run 'python check.py' to isolate which component is failing.
No audio input detected — microphone not working
Check that your microphone is recognized by the OS and that FunASR is running and accessible at the URL set in kuonasr/config.yaml. You can disable audio input by setting audio_input_sw: false in the root config.yaml and use text-only mode instead.
pip install fails for torch or audio dependencies
If you don't need offline GPU-accelerated models, skip the PyTorch install — it's only needed for local offline models. Run 'pip install -r requirements.txt' without the torch install command if you're using API-based speech recognition and TTS.
Frequently Asked Questions about Kuon
What is Kuon?
Kuon is a Model Context Protocol (MCP) server that 久远：一个开发中的大模型语音助手，当前关注易用性，简单上手，支持对话选择性记忆和model context protocol (mcp)服务。 kuon:a large language model-based voice assistant under development, currently focused on ease of use and simple onboarding. it supports selective memory in conversations and the It connects AI assistants to external tools and data sources through a standardized interface.
How do I install Kuon?
Follow the installation instructions on the Kuon GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with Kuon?
Kuon works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is Kuon free to use?
Yes, Kuon is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
Kuon Alternatives — Similar Communication Servers
Looking for alternatives to Kuon? Here are other popular communication servers you can use with Claude, Cursor, and VS Code.
Slack
★ 86.1kEnables interaction with Slack workspaces through comprehensive channel management, messaging, user management, file uploads, and Block Kit formatting. Features secure credential storage via macOS Keychain and supports all major Slack operations incl
LibreChat
★ 37.3kEnhanced ChatGPT Clone: Features Agents, MCP, DeepSeek, Anthropic, AWS, OpenAI, Responses API, Azure, Groq, o1, GPT-5, Mistral, OpenRouter, Vertex AI, Gemini, Artifacts, AI model switching, message search, Code Interpreter, langchain, DALL-E-3, OpenA
Note Gen
★ 11.9kA cross-platform Markdown AI note-taking software.
DeepChat
★ 5.8k🐬DeepChat - A smart assistant that connects powerful AI to your personal world
WhatsApp MCP
★ 5.7kConnects AI assistants to the official Meta WhatsApp Cloud API for managing conversations and sending various message types through natural language. It provides tools for media management, template messages, and real-time webhook processing without
Apple
★ 3.1kEnables interaction with Apple apps like Messages, Notes, and Contacts through the MCP protocol to send messages, search, and open app content using natural language.
Browse More Communication MCP Servers
Explore all communication servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up Kuon in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use Kuon?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.