MCP Client for Ollama
A text-based user interface (TUI) client for interacting with MCP servers using Ollama. Features include agent mode, multi-server, model switching, streaming responses, tool management, human-in-the-loop, thinking mode, model params config, MCP promp
What is MCP Client for Ollama?
MCP Client for Ollama is a Model Context Protocol (MCP) server that allows AI assistants like Claude, Cursor, and VS Code to text-based user interface (tui) client for interacting with mcp servers using ollama. features include agent mode, multi-server, model switching, streaming responses, tool management, human-in-the-loo...
A text-based user interface (TUI) client for interacting with MCP servers using Ollama. Features include agent mode, multi-server, model switching, streaming responses, tool management, human-in-the-loop, thinking mode, model params config, MCP promp
This server falls under the Developer Tools category on MCPgee, the world's largest MCP server directory with 33,000+ servers.
Features
- A text-based user interface (TUI) client for interacting wit
Use Cases
Maintainer
Works with
Installation
Manual Installation
npx mcp-client-for-ollamaConfiguration
Configuration Details
claude_desktop_config.json
Performance
Response Metrics
Resource Usage
How to Set Up and Use MCP Client for Ollama
MCP Client for Ollama (ollmcp) is a terminal-based TUI application that lets you interact with any MCP server using locally running Ollama models, bringing MCP tool-calling to fully offline and privacy-preserving workflows. It supports STDIO, SSE, and Streamable HTTP transports, connects to multiple MCP servers simultaneously, provides an agent mode with iterative tool execution, human-in-the-loop approval for tool calls, model parameter tuning, vision support, conversation history export, and live hot-reload of MCP servers without restarting the client.
Prerequisites
- Python 3.10 or higher installed
- Ollama installed and running locally at http://localhost:11434 (ollama.ai)
- At least one Ollama model pulled that supports tool calling (e.g. qwen2.5:7b)
- pip or uv package manager for installation
- One or more MCP servers to connect to (can auto-discover from Claude Desktop config)
Install ollmcp via pip or uv
Install the mcp-client-for-ollama package (published as ollmcp on PyPI). The uv method runs without a permanent install using uvx.
# Install with pip:
pip install --upgrade ollmcp
# Or run directly with uv (no install needed):
uvx ollmcpEnsure Ollama is running with a tool-capable model
Start Ollama and pull a model that supports function/tool calling. The default model is qwen2.5:7b, which works well with MCP tools.
ollama serve
ollama pull qwen2.5:7bLaunch ollmcp with auto-discovery
Use the --auto-discovery flag to automatically load all MCP servers defined in your Claude Desktop config. This is the fastest way to get started if you already have MCP servers configured.
ollmcp --auto-discoveryOr connect to a specific MCP server
Connect to a specific MCP server by script path, URL, or JSON config file. Use -m to select a different Ollama model.
# Connect to a local MCP server script:
ollmcp -s /path/to/server.py -m llama3.2:3b
# Connect to an SSE server:
ollmcp -u http://localhost:8080/sse
# Load from a JSON config:
ollmcp -j ~/.config/mcp-servers.jsonUse interactive commands inside the TUI
Once running, use slash commands to switch models, manage tools, enable human-in-the-loop approval, and export conversation history.
/model # Switch to a different Ollama model
/tools # Enable or disable specific MCP tools
/human-in-the-loop # Toggle tool execution approval
/agent # Toggle agent mode for iterative tasks
/export-history # Save the conversation to JSON
/reload-servers # Hot-reload MCP servers without restartMCP Client for Ollama Examples
Client configuration
ollmcp is itself an MCP client, not a server. Configure it with a servers JSON file to connect multiple MCP servers simultaneously. Example ~/.config/mcp-servers.json layout:
{
"mcpServers": {
"filesystem": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-filesystem", "/tmp"]
},
"web-search": {
"command": "npx",
"args": ["-y", "@modelcontextprotocol/server-brave-search"],
"env": {
"BRAVE_API_KEY": "your-brave-api-key"
}
}
}
}Prompts to try
Sample tasks that use ollmcp's agent mode and MCP tool integrations with local Ollama models.
- "Search the web for the latest news on MCP protocol adoption and summarize the top 3 results."
- "List the files in /tmp, find any log files, and show me the last 20 lines of the most recent one."
- "In agent mode, research Python async best practices and write a cheat sheet to ~/async-tips.md."Troubleshooting MCP Client for Ollama
Connection refused to Ollama at localhost:11434
Ensure Ollama is running with `ollama serve`. If Ollama is on a different host or port, use the -H flag: `ollmcp -H http://192.168.1.100:11434`. Verify with `curl http://localhost:11434/api/tags`.
Model does not call MCP tools and just answers from training data
Not all Ollama models support tool calling. Switch to a verified tool-capable model like qwen2.5:7b, llama3.1:8b, or mistral-nemo using `/model` inside the TUI or `-m model-name` on startup.
Auto-discovery finds no MCP servers
Auto-discovery reads from ~/Library/Application Support/Claude/claude_desktop_config.json (macOS). If your Claude Desktop config is empty or in a non-standard location, use `-j path/to/config.json` to provide a custom servers config file instead.
Frequently Asked Questions about MCP Client for Ollama
What is MCP Client for Ollama?
MCP Client for Ollama is a Model Context Protocol (MCP) server that text-based user interface (tui) client for interacting with mcp servers using ollama. features include agent mode, multi-server, model switching, streaming responses, tool management, human-in-the-loop, thinking mode, model params config, mcp promp It connects AI assistants to external tools and data sources through a standardized interface.
How do I install MCP Client for Ollama?
Follow the installation instructions on the MCP Client for Ollama GitHub repository. Clone the repo, install dependencies, and add the server config to your AI client.
Which AI clients work with MCP Client for Ollama?
MCP Client for Ollama works with all major MCP-compatible AI clients including Claude Desktop, Claude Code, Cursor, VS Code (GitHub Copilot), Windsurf, and Cline.
Is MCP Client for Ollama free to use?
Yes, MCP Client for Ollama is open source and available under the MIT license. You can use it freely in both personal and commercial projects.
MCP Client for Ollama Alternatives — Similar Developer Tools Servers
Looking for alternatives to MCP Client for Ollama? Here are other popular developer tools servers you can use with Claude, Cursor, and VS Code.
Ecc
★ 188.2kThe agent harness performance optimization system. Skills, instincts, memory, security, and research-first development for Claude Code, Codex, Opencode, Cursor and beyond.
Javaguide
★ 155.8kJava 面试 & 后端通用面试指南，覆盖计算机基础、数据库、分布式、高并发、系统设计与 AI 应用开发
Gemini CLI
★ 104.5kA secure MCP server that wraps the Google Gemini CLI, allowing clients to query Gemini models using local OAuth sessions without requiring an API key. It provides tools for model interaction and diagnostics with built-in protection against command in
Awesome MCP Servers
★ 87.3k⭐ Curated list of Model Context Protocol (MCP) servers - tools that extend Claude Desktop, Cursor, Windsurf, and other MCP clients with custom capabilities.
MCP Servers
★ 86.0kModel Context Protocol Servers
CC Switch
★ 77.5kA cross-platform desktop All-in-One assistant for Claude Code, Codex, OpenCode, OpenClaw, Gemini CLI & Hermes Agent. Only official website: ccswitch.io
Browse More Developer Tools MCP Servers
Explore all developer tools servers available in the MCPgee directory. Each server includes setup guides for Claude, Cursor, and VS Code.
Set Up MCP Client for Ollama in Your Editor
Choose your AI client for step-by-step setup instructions.
Quick Config Preview
Add this to your claude_desktop_config.json or .cursor/mcp.json
Ready to use MCP Client for Ollama?
Browse our complete directory of 33,000+ MCP servers, read setup guides for your editor, and start building with the Model Context Protocol.