Skip to main content
Glama

LibreModel MCP Server ๐Ÿค–

A Model Context Protocol (MCP) server that bridges Claude Desktop with your local LLM instance running via llama-server.

Features

  • ๐Ÿ’ฌ Full conversation support with Local Model through Claude Desktop

  • ๐ŸŽ›๏ธ Complete parameter control (temperature, max_tokens, top_p, top_k)

  • โœ… Health monitoring and server status checks

  • ๐Ÿงช Built-in testing tools for different capabilities

  • ๐Ÿ“Š Performance metrics and token usage tracking

  • ๐Ÿ”ง Easy configuration via environment variables

Related MCP server: Llama 4 Maverick MCP Server

Quick Start

npm install @openconstruct/llama-mcp-server

A Model Context Protocol (MCP) server that bridges Claude Desktop with your local LLM instance running via llama-server.

Features

  • ๐Ÿ’ฌ Full conversation support with LibreModel through Claude Desktop

  • ๐ŸŽ›๏ธ Complete parameter control (temperature, max_tokens, top_p, top_k)

  • โœ… Health monitoring and server status checks

  • ๐Ÿงช Built-in testing tools for different capabilities

  • ๐Ÿ“Š Performance metrics and token usage tracking

  • ๐Ÿ”ง Easy configuration via environment variables

Quick Start

1. Install Dependencies

cd llama-mcp
npm install

2. Build the Server

npm run build

3. Start Your LibreModel

Make sure llama-server is running with your model:

./llama-server -m lm37.gguf -c 2048 --port 8080

4. Configure Claude Desktop

Add this to your Claude Desktop configuration (~/.config/claude/claude_desktop_config.json):

{
  "mcpServers": {
    "libremodel": {
      "command": "node",
      "args": ["/home/jerr/llama-mcp/dist/index.js"]
    }
  }
}

5. Restart Claude Desktop

Claude will now have access to LibreModel through MCP!

Usage

Once configured, you can use these tools in Claude Desktop:

๐Ÿ’ฌ chat - Main conversation tool

Use the chat tool to ask LibreModel: "What is your name and what can you do?"

๐Ÿงช quick_test - Test LibreModel capabilities

Run a quick_test with type "creative" to see if LibreModel can write poetry

๐Ÿฅ health_check - Monitor server status

Use health_check to see if LibreModel is running properly

Configuration

Set environment variables to customize behavior:

export LLAMA_SERVER_URL="http://localhost:8080"  # Default llama-server URL

Available Tools

Tool

Description

Parameters

chat

Converse with MOdel

message, temperature, max_tokens, top_p, top_k, system_prompt

quick_test

Run predefined capability tests

test_type (hello/math/creative/knowledge)

health_check

Check server health and status

None

Resources

  • Configuration: View current server settings

  • Instructions: Detailed usage guide and setup instructions

Development

# Install dependencies
npm install # LibreModel MCP Server ๐Ÿค–

A Model Context Protocol (MCP) server that bridges Claude Desktop with your local LLM instance running via llama-server.

## Features

- ๐Ÿ’ฌ **Full conversation support** with LibreModel through Claude Desktop
- ๐ŸŽ›๏ธ **Complete parameter control** (temperature, max_tokens, top_p, top_k)
- โœ… **Health monitoring** and server status checks
- ๐Ÿงช **Built-in testing tools** for different capabilities
- ๐Ÿ“Š **Performance metrics** and token usage tracking
- ๐Ÿ”ง **Easy configuration** via environment variables

## Quick Start

### 1. Install Dependencies

```bash
cd llama-mcp
npm install

2. Build the Server

npm run build

3. Start Your LibreModel

Make sure llama-server is running with your model:

./llama-server -m lm37.gguf -c 2048 --port 8080

4. Configure Claude Desktop

Add this to your Claude Desktop configuration (~/.config/claude/claude_desktop_config.json):

{
  "mcpServers": {
    "libremodel": {
      "command": "node",
      "args": ["/home/jerr/llama-mcp/dist/index.js"]
    }
  }
}

5. Restart Claude Desktop

Claude will now have access to LibreModel through MCP!

Usage

Once configured, you can use these tools in Claude Desktop:

๐Ÿ’ฌ chat - Main conversation tool

Use the chat tool to ask LibreModel: "What is your name and what can you do?"

๐Ÿงช quick_test - Test LibreModel capabilities

Run a quick_test with type "creative" to see if LibreModel can write poetry

๐Ÿฅ health_check - Monitor server status

Use health_check to see if LibreModel is running properly

Configuration

Set environment variables to customize behavior:

export LLAMA_SERVER_URL="http://localhost:8080"  # Default llama-server URL

Available Tools

Tool

Description

Parameters

chat

Converse with MOdel

message, temperature, max_tokens, top_p, top_k, system_prompt

quick_test

Run predefined capability tests

test_type (hello/math/creative/knowledge)

health_check

Check server health and status

None

Resources

  • Configuration: View current server settings

  • Instructions: Detailed usage guide and setup instructions

Development

# Install dependencies
npm install openconstruct/llama-mcp-server


# Development mode (auto-rebuild)
npm run dev

# Build for production
npm run build

# Start the server directly
npm start

Architecture

Claude Desktop โ†โ†’ LLama MCP Server โ†โ†’ llama-server API โ†โ†’ Local Model

The MCP server acts as a bridge, translating MCP protocol messages into llama-server API calls and formatting responses for Claude Desktop.

Troubleshooting

"Cannot reach LLama server"

  • Ensure llama-server is running on the configured port

  • Check that the model is loaded and responding

  • Verify firewall/network settings

"Tool not found in Claude Desktop"

  • Restart Claude Desktop after configuration changes

  • Check that the path to index.js is correct and absolute

  • Verify the MCP server builds without errors

Poor response quality

  • Adjust temperature and sampling parameters

  • Try different system prompts

License

CC0-1.0 - Public Domain. Use freely!


Built with โค๏ธ for open-source AI and the LibreModel project. by Claude Sonnet4

Development mode (auto-rebuild)

npm run dev

Build for production

npm run build

Start the server directly

npm start


## Architecture

Claude Desktop โ†โ†’ LLama MCP Server โ†โ†’ llama-server API โ†โ†’ Local Model


The MCP server acts as a bridge, translating MCP protocol messages into llama-server API calls and formatting responses for Claude Desktop.

## Troubleshooting

**"Cannot reach LLama server"**
- Ensure llama-server is running on the configured port
- Check that the model is loaded and responding
- Verify firewall/network settings

**"Tool not found in Claude Desktop"**
- Restart Claude Desktop after configuration changes
- Check that the path to `index.js` is correct and absolute
- Verify the MCP server builds without errors

**Poor response quality**
- Adjust temperature and sampling parameters
- Try different system prompts

## License

CC0-1.0 - Public Domain. Use freely!

---

Built with โค๏ธ for open-source AI and the LibreModel project. by Claude Sonnet4

### 1. Install Dependencies

```bash
cd llama-mcp
npm install

2. Build the Server

npm run build

3. Start Your LibreModel

Make sure llama-server is running with your model:

./llama-server -m lm37.gguf -c 2048 --port 8080

4. Configure Claude Desktop

Add this to your Claude Desktop configuration (~/.config/claude/claude_desktop_config.json):

{
  "mcpServers": {
    "libremodel": {
      "command": "node",
      "args": ["/home/jerr/llama-mcp/dist/index.js"]
    }
  }
}

5. Restart Claude Desktop

Claude will now have access to LibreModel through MCP!

Usage

Once configured, you can use these tools in Claude Desktop:

๐Ÿ’ฌ chat - Main conversation tool

Use the chat tool to ask LibreModel: "What is your name and what can you do?"

๐Ÿงช quick_test - Test LibreModel capabilities

Run a quick_test with type "creative" to see if LibreModel can write poetry

๐Ÿฅ health_check - Monitor server status

Use health_check to see if LibreModel is running properly

Configuration

Set environment variables to customize behavior:

export LLAMA_SERVER_URL="http://localhost:8080"  # Default llama-server URL

Available Tools

Tool

Description

Parameters

chat

Converse with MOdel

message, temperature, max_tokens, top_p, top_k, system_prompt

quick_test

Run predefined capability tests

test_type (hello/math/creative/knowledge)

health_check

Check server health and status

None

Resources

  • Configuration: View current server settings

  • Instructions: Detailed usage guide and setup instructions

Development

# Install dependencies
npm install

# Development mode (auto-rebuild)
npm run dev

# Build for production
npm run build

# Start the server directly
npm start

Architecture

Claude Desktop โ†โ†’ LLama MCP Server โ†โ†’ llama-server API โ†โ†’ Local Model

The MCP server acts as a bridge, translating MCP protocol messages into llama-server API calls and formatting responses for Claude Desktop.

Troubleshooting

"Cannot reach LLama server"

  • Ensure llama-server is running on the configured port

  • Check that the model is loaded and responding

  • Verify firewall/network settings

"Tool not found in Claude Desktop"

  • Restart Claude Desktop after configuration changes

  • Check that the path to index.js is correct and absolute

  • Verify the MCP server builds without errors

Poor response quality

  • Adjust temperature and sampling parameters

  • Try different system prompts

License

CC0-1.0 - Public Domain. Use freely!


Built with โค๏ธ for open-source AI and the LibreModel project. by Claude Sonnet4

Available Tools

3 tools
chatChat with LibreModelC

Have a conversation with LibreModel (Gigi)

Inputs: message (required), temperature, max_tokens, top_p, +2 more.

health_checkCheck LibreModel Server HealthB

Check if the llama-server is running and responsive

Inputs: no named parameters.

quick_testQuick LibreModel TestB

Run a quick test to see if LibreModel is responding

Inputs: test_type.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelogโ€ฆ

    TDQS

    C2.8/5.0

    Scored across 3 tools

    Disambiguation2/5

    The tools have overlapping purposes that could cause confusion. 'chat' and 'quick_test' both involve interacting with LibreModel to check responsiveness, making them ambiguous. 'health_check' is more distinct but still overlaps in verifying server status.

    Naming Consistency4/5

    The naming follows a consistent snake_case pattern across all tools, which is clear and readable. However, the verb styles are mixed ('chat' is a noun-like verb, 'health_check' and 'quick_test' are noun-based), slightly reducing consistency.

    Tool Count2/5

    With only 3 tools, the set feels thin for a model server, lacking essential operations like model loading, configuration, or inference control. This minimal count suggests an incomplete surface for the domain.

    Completeness1/5

    The tool set is severely incomplete for a model server. There are no tools for core functions such as loading models, managing sessions, adjusting parameters, or handling inputs/outputs beyond basic chat and health checks, leaving significant gaps.

    Maintenance

    ActivityInactive
    ResponsivenessOutdated data

    Related MCP Connectors

    Related MCP Servers

    • A
      license
      Not graded
      quality
      B
      maintenance
      A bridge that allows Claude to communicate with locally running LLM models via LM Studio, enabling users to leverage their private models through Claude's interface.
      173
      MIT
    • A
      license
      Not graded
      quality
      D
      maintenance
      Bridges Llama models with Claude Desktop through Ollama, enabling privacy-first local AI operations with 10+ built-in tools for file operations, web search, calculations, and custom model deployment. Features streaming support, hybrid intelligence workflows, and extensive Python ecosystem integration for research, development, and enterprise applications.
      MIT
    • A
      license
      A
      quality
      D
      maintenance
      A bridge that enables Claude Code to interact with local Ollama instances for text generation, multi-turn chat, and vision-based analysis. It supports model management tasks such as listing, pulling, and showing details, alongside generating text embeddings.
      6
      1
      MIT