MCP Client for Ollama (ollmcp)
A terminal MCP client for local Ollama models: MCP tools, prompts and resources, an agent mode and model switching
Medium risk
We rate an entry medium when the tool runs code, makes network calls or reads project files. Check what exactly it does before installing.
Why this level
- Runs tools from connected MCP servers
- Can execute model-suggested actions, with human confirmation available
Install
In your terminal, with SkillFoxx CLI
npx skillfoxx add cli/mcp-client-for-ollama-ollmcpDetects the agents on your machine, checks the risk and pins the version.
Other ways to install
Install the tool
pipx install mcp-client-for-ollamaWithout pipx, pip install mcp-client-for-ollama works too.
Install via uv: uv tool install --upgrade ollmcp, or pip install --upgrade ollmcp. Add a server with ollmcp mcp add, then run ollmcp.
Other ways from the author
uv tool install --upgrade ollmcpOr pip install --upgrade ollmcp. Run it with ollmcp.
This is third-party code. Review the repository files before installing.
What it does
ollmcp is a terminal client that lets local Ollama models work with MCP servers and run tools. It supports all core MCP primitives: tools, prompts and resources, plus an agent mode, connecting several servers at once, model switching, streaming responses, human-in-the-loop confirmation, a thinking mode, model parameter config, system prompts and saved preferences. The stdio, SSE and streamable HTTP transports let you connect different servers. It installs from PyPI as ollmcp, servers are added with ollmcp mcp add, and the interface is a text UI with commands like /help. It runs on Windows, macOS and Linux.
Who it is for. For developers who want to connect MCP servers to local Ollama models straight from the terminal.
Good fit when
- You work with local models via Ollama and want MCP tools
- You need a client with an agent mode and action confirmation
- You need to connect several MCP servers at once
Not a fit when
- You do not use Ollama and local models
- You need a graphical interface, not a terminal
Example request
Connect the playwright MCP server and let the local model run browser actionsLimitations
The client targets Ollama and local models and is not meant for cloud models. It needs Python 3.11 and Ollama installed with a model pulled. Tool capabilities depend on the connected MCP servers and on how well the model calls tools.
How to disable. Remove the ollmcp package: uv tool uninstall ollmcp or pip uninstall ollmcp.
Security check
- Runs tools from connected MCP servers
- Can execute model-suggested actions, with human confirmation available
README in short
The README describes ollmcp as a terminal MCP client for local Ollama models with the full set of primitives: tools, prompts and resources. It has an agent mode, multi-server connections, model switching, streaming responses, action confirmation, a thinking mode and saved preferences. The stdio, SSE and streamable HTTP transports are supported. Install is via uv or pip as ollmcp, and servers are added with ollmcp mcp add. It runs on Windows, macOS and Linux. MIT licensed.
FAQ
Which transports are supported?
stdio, SSE and streamable HTTP, and you can connect several servers at once.
Do I need cloud access?
No, the client works with local models via Ollama.
Related
Open-source personal AI assistant on your own machine: answers in Telegram, Slack, Discord and WhatsApp, extended with skills and plugins
A self-improving agent from Nous Research with a TUI, messaging gateway, cron jobs and skills it writes itself
An open source coding agent for the terminal and desktop with build and plan modes
Open prompt library with a Claude Code plugin, MCP server and CLI: search, fetch and improve prompts and skills from an agent