MCP Ollama Server
# MCP Ollama
A Model Context Protocol (MCP) server for integrating Ollama with Claude Desktop or other MCP clients.
<a href="https://glama.ai/mcp/servers/h0t3210s62"><img width="380" height="200" src="https://glama.ai/mcp/servers/h0t3210s62/badge" alt="Ollama Server MCP server" /></a>
## Requirements
- Python 3.10 or higher
- Ollama installed and running (https://ollama.com/download)
- At least one model pulled with Ollama (e.g., `ollama pull llama2`)
### Configure Claude Desktop
Add to your Claude Desktop configuration (`~/Library/Application Support/Claude/claude_desktop_config.json` on macOS, `%APPDATA%\Claude\claude_desktop_config.json` on Windows):
```json
{
"mcpServers": {
"ollama": {
"command": "uvx",
"args": [
"mcp-ollama"
]
}
}
}
```
### Development
Install in development mode:
```bash
git clone https://github.com/yourusername/mcp-ollama.git
cd mcp-ollama
uv sync
```
Test with MCP Inspector:
```bash
mcp dev src/mcp_ollama/server.py
```
## Features
The server provides four main tools:
- `list_models` - List all downloaded Ollama models
- `show_model` - Get detailed information about a specific model
- `ask_model` - Ask a question to a specified model
## License
MIT
TDQS
Scored across 3 tools
Each tool has a clearly distinct purpose with no overlap. 'ask_model' queries a model, 'list_models' enumerates available models, and 'show_model' provides metadata about a specific model. An agent can easily distinguish between these three distinct operations.
All tool names follow a consistent verb_noun pattern using snake_case. 'ask_model', 'list_models', and 'show_model' maintain perfect naming consistency throughout the set, making them predictable and readable.
With only 3 tools, the server feels somewhat thin for interacting with Ollama models. While the tools cover basic operations, additional functionality like model management (e.g., pull, delete) or conversation history would be expected for a more complete surface.
The tools provide good coverage for core Ollama interactions: querying models, listing them, and getting details. However, there are minor gaps in model management operations (e.g., pulling new models, deleting models) that agents might need to work around for full lifecycle coverage.