Ollama MCP Server (Daniel Steigman)
Ollama local LLM management and chat completion via the Model Context Protocol.
Data last scanned yesterday · Reviewed last month
Overview
This MCP server provides a bridge between the Ollama local model ecosystem and MCP-compatible clients. It enables users to pull, push, and list local models, as well as create custom instances from Modelfiles, directly through an interface that mirrors the OpenAI chat completion API. Developers can manage Ollama server settings and execute complex prompts with configurable parameters such as temperature and timeout. The server is designed to integrate into environments like Claude Desktop, allowing for granular control over model execution and local model registry management.
Our verdict
With a security score of 10/100, this server carries a high-risk rating. This score reflects significant maintenance concerns, specifically that the last commit was 573 days ago and the project lacks an explicit open-source license. Given the lack of active updates and unverified security status, this tool should be treated as experimental and is not recommended for production or sensitive environments.
- Supports full Ollama API management and model lifecycle operations
- Provides an OpenAI-compatible interface for chat completion
- Features a small footprint with only two direct dependencies
- Stagnant development with no commits in over 570 days
- No open-source license provided for legal compliance
- Community-maintained project with single-contributor bus-factor risk
Setup
{
"mcpServers": {
"nighttrek-ollama-mcp": {
"args": [
"-y",
"ollama-mcp"
],
"command": "npx"
}
}
}Paste into ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows), then restart Claude Desktop. Some servers require API keys or tokens as environment variables — check the project's README and add an "env" object if needed.
- No read-only mode — running this grants full read/write access.
- No built-in authentication — anyone with access to this config can invoke every tool it exposes.
Tools
| Tool | Description | Risk |
|---|---|---|
| pull | Pull a model from a registry. | medium |
| run | Run a model with a prompt. | low |
| chat_completion | Perform chat completion using an OpenAI-compatible interface. | low |
| create | Create a custom model from a Modelfile. | medium |
Compatibility
| Client | Local | Docker | Remote | Read-only |
|---|---|---|---|---|
| ChatGPT | ||||
| Claude Desktop | ||||
| Cursor | ||||
| VS Code | ||||
| Windsurf |
Frequently asked questions
›What are the software requirements to run this server?
You must have Node.js, npm or pnpm, and Ollama installed on your system.
›How can I connect the server to a custom Ollama API endpoint?
You can specify a custom endpoint by setting the OLLAMA_HOST environment variable in your MCP configuration file.
›Does this server support custom model creation?
Yes, you can use the create tool to build custom models by providing a name and a path to a Modelfile.
›Can I adjust the randomness of the model responses?
Yes, you can configure the temperature parameter within the chat completion tool, which accepts values in the range of 0 to 2.
›Is the chat completion tool compatible with standard APIs?
The chat completion tool is designed as a drop-in replacement for the OpenAI chat completion API.
Alternatives
Badge
Maintain this server? Add the live badge to your README.