← Discover MCPs and Agents
L
MCPAI & MLGitHub

LLMule-client

Links

README

From the repo.

LLMule Client

A peer-to-peer client for sharing Large Language Models (LLMs) across the LLMule network. Run your local LLMs and share them with the community.

🌐 Official Website: https://llmule.xyz 💬 Join our Community: Discord Channel

Features

  • Automatic detection of local LLM models (Ollama & LM Studio)
  • Real-time connection to the LLMule network
  • Model tier categorization (Tiny, Small, Medium)
  • Health monitoring and automatic reconnection
  • Secure API key authentication

Prerequisites

  • Node.js v20 or higher
  • One of the following LLM backends:

Installation

  1. Clone the repository:
git clone https://github.com/cm64-studio/LLMule-client.git
cd LLMule-client
  1. Install dependencies:
npm install
  1. Create a configuration file:
cp .env.example .env

Configuration

Edit .env file with your settings. Don't worry about the API key - you'll get it automatically during the first run registration process at llmule.xyz.

# Server Configuration
API_URL=https://api.llmule.xyz
SERVER_URL=wss://api.llmule.xyz/llm-network

# LLM Provider URLs (defaults)
OLLAMA_URL=http://localhost:11434
LMSTUDIO_URL=http://localhost:1234/v1

# Advanced
LOG_LEVEL=info
MAX_RETRIES=5

Supported Models

Tier 1 - Small (3B)

  • TinyLlama
  • Minimum Requirements: 4GB RAM

Tier 2 - Medium (7B)

  • Mistral 7B
  • Minimum Requirements: 8GB RAM

Tier 3 - Large (14B)

  • Microsoft Phi-4
  • Minimum Requirements: 16GB RAM

Supported LLM Providers

LLMule supports the following LLM providers:

  • Ollama: Run models like Llama, Mistral, and more locally
  • LM Studio: Run various open-source models with a nice UI
  • EXO: Run distributed models across multiple devices

Configuration

Set up your providers in your .env file:

OLLAMA_URL=http://localhost:11434
LMSTUDIO_URL=http://localhost:1234/v1
EXO_URL=http://localhost:52415

Usage

  1. Start your LLM backend (Ollama or LM Studio)

  2. Run the client:

npm start
  1. First-time setup:
    • On first run, you'll be guided through the registration process at llmule.xyz
    • Your API key will be automatically configured after registration
    • Select the models you want to share
    • The client will automatically connect to the LLMule network

Running as a Service

Systemd Service (Linux)

  1. Create service file:
sudo nano /etc/systemd/system/llmule-client.service
  1. Add configuration:
[Unit]
Description=LLMule Client
After=network.target ollama.service

[Service]
Type=simple
User=ubuntu
WorkingDirectory=/path/to/llmule-client
ExecStart=/usr/bin/npm start
Restart=always
Environment=NODE_ENV=production

[Install]
WantedBy=multi-user.target
  1. Enable and start:
sudo systemctl enable llmule-client
sudo systemctl start llmule-client

Monitoring

Check client status:

npm run status

View logs:

# Live logs
npm run logs

# Error logs
npm run logs:error

Troubleshooting

Common issues and solutions:

  1. Connection Issues
# Check network connectivity
curl -v $SERVER_URL

# Verify Ollama is running
curl http://localhost:11434/api/tags
  1. Model Detection Issues
# List Ollama models
ollama list

# Check LM Studio API
curl http://localhost:1234/v1/models

Contributing

  1. Fork the repository
  2. Create feature branch
  3. Commit changes
  4. Push to branch
  5. Create Pull Request

Security

  • API keys are stored securely
  • All network traffic is encrypted
  • Models are sandboxed
  • Resource limits enforced

Support

License

MIT License - see LICENSE for details

Collected info

  • 6 stars
  • Language: JavaScript
  • Source updated: 6/15/2026

Config for your environment

Replace {MCP_ENDPOINT_URL} with this MCP’s endpoint URL (from its repo or docs above). No API key — you connect directly.

Tool

OS

Config file: ~/.cursor/mcp.json

{
  "mcpServers": {
    "mcp-server": {
      "url": "{MCP_ENDPOINT_URL}"
    }
  }
}

Paste into mcpServers in the config file. Restart Cursor after saving.

If this MCP is also published on mcpchannel.ai, you can subscribe from Browse and use the gateway config there instead.