🚀 MCP-Ollama Server
About
Extends Model Context Protocol (MCP) to local LLMs via Ollama, enabling Claude-like tool use (files, web, email, GitHub, AI images) while keeping data private. Modular Python servers for on-prem AI. #LocalAI #MCP #Ollama
Details
- License
- Apache-2.0
Explore
- 🔒 Complete Data Privacy: All computations happen locally through Ollama
- 🔧 Tool Use for Local LLMs: Extends Ollama models with file, calendar, and other capabilities
- 🧩 Modular Architecture: Independent Python service modules that can be deployed selectively
- 🔌 Easy Integration: Simple APIs to connect with existing applications
- 🚀 Performance Optimized: Minimal overhead to maintain responsive AI interactions
- 📦 Containerized Deployment: Docker support for each module (coming soon)
- 🧪 Extensive Testing: Comprehensive test coverage for reliability
Setting up with Highlight
This MCP is not yet compatible with Highlight’s one-click setup. However, you can still use it with Highlight by following these steps:
- Download and install Highlight from highlightai.com/download
- Navigate to the plugins tab and select "Add Custom Plugin"
-
Configure the plugin with the settings below
Plugin Name
🚀 MCP-Ollama ServerCommand (node, npx, python, etc.)Please refer to the README for specific instructions on how to obtain API keys or other required environment variables.
- Enable "Start Automatically" if you want the plugin to start when Highlight launches
From the repository
- Python 3.8+ installed
- Ollama set up on your system
- Git for cloning the repository
curl -LsSf https://astral.sh/uv/install.sh | sh
ollama pull llama3
1. 📅 Calendar Module:
cd calendar
uv add pyproject.toml # Install calendar-specific dependencies
2. 🔄 Client MCP Module:
cd client_mcp
uv add pyproject.toml # Install calendar-specific dependencies
3. 📁 File System Module:
cd file_system
uv add pyproject.toml # Install filesystem dependencies
```
Claude Desktop / Cursor
Paste into your MCP client config file to install this server.
{
"mcpServers": {
"\ud83d\ude80 mcp-ollama server": {
"MCP-ollama_server": {
"command": "uv",
"args": [
"run",
"client.py",
"../file_system/file_system.py"
]
}
}
}
}McpServers
{
"MCP-ollama_server": {
"command": "uv",
"args": [
"run",
"client.py",
"../file_system/file_system.py"
]
}
}<div align="center">
Connect the power of Model Context Protocol with local LLMs
Getting Started •
Features •
Architecture •
Documentation •
Contributing •
FAQ
</div>
📋 Overview
MCP-Ollama Server bridges the gap between Anthropic's Model Context Protocol (MCP) and local LLMs via Ollama. This integration empowers your on-premise AI models with Claude-like tool capabilities, including file system access, calendar integration, web browsing, email communication, GitHub interactions, and AI image generation—all while maintaining complete data privacy.
Unlike cloud-based AI solutions, MCP-Ollama Server:
- Keeps all data processing on your local infrastructure
- Eliminates the need to share sensitive information with third parties
- Provides a modular approach that allows you to use only the components you need
- Enables enterprise-grade AI capabilities in air-gapped or high-security environments
✨ Key Features
- 🔒 Complete Data Privacy: All computations happen locally through Ollama
- 🔧 Tool Use for Local LLMs: Extends Ollama models with file, calendar, and other capabilities
- 🧩 Modular Architecture: Independent Python service modules that can be deployed selectively
- 🔌 Easy Integration: Simple APIs to connect with existing applications
- 🚀 Performance Optimized: Minimal overhead to maintain responsive AI interactions
- 📦 Containerized Deployment: Docker support for each module (coming soon)
- 🧪 Extensive Testing: Comprehensive test coverage for reliability
🚀 Quick Start
Prerequisites
- Python 3.8+ installed
- Ollama set up on your system
- Git for cloning the repository
🧩 Component Overview
MCP-Ollama Server is organized into specialized modules, each providing specific functionality:
📅 Calendar Module
calendar/
├── README.md # Module-specific documentation
├── google_calendar.py # Google Calendar API integration
├── pyproject.toml # Dependencies and package info
└── uv.lock # Dependency lock file
The Calendar module enables your local LLM to:
- Create, modify, and delete calendar events
- Check availability and scheduling conflicts
- Send meeting invitations
- Set reminders and notifications
🔄 Client MCP Module
client_mcp/
├── README.md # Module-specific documentation
├── client.py # Main client implementation
├── pyproject.toml # Dependencies and package info
├── testing.txt # Test data
└── uv.lock # Dependency lock file
The Client module provides:
- A unified interface to interact with all MCP-enabled services
- Conversation history management
- Context handling for improved responses
- Tool selection and routing logic
📁 File System Module
file_system/
├── README.md # Module-specific documentation
├── file_system.py # File system operations implementation
├── pyproject.toml # Dependencies and package info
└── uv.lock # Dependency lock file
The File System module allows your local LLM to:
- Read and write files securely
- List directory contents
- Search for files matching specific patterns
- Parse different file formats (text, CSV, JSON, etc.)
Installation
```bash
Sign in to leave a review
Use Google, GitHub, or an email account so ratings stay tied to real people.
No reviews posted yet.
