Ollama Bridge MCP Server
Local Ollama models execute tools through diverse MCP-compatible server plugins.
Data last scanned today · Reviewed yesterday
Overview
This TypeScript bridge acts as a translation layer between local Large Language Models hosted on Ollama and various MCP servers. By managing tool registration, JSON-RPC communication, and dynamic request routing, it enables models like Qwen 2.5 to leverage external tools for filesystem access, web searches, and third-party integrations.
Our verdict
With a security score of 13/100, this project carries a high risk level. The lack of any built-in authentication means anyone with access to the configuration file can execute tool calls, and the absence of a read-only mode grants the bridge full write capabilities. Given the project has not been updated in over 450 days, it is recommended only for local testing in isolated environments rather than sensitive workflows.
- Supports integration with Brave Search, GitHub, and Gmail.
- Small dependency footprint with only 4 direct packages.
- Includes documentation for multi-server MCP configurations.
- No authentication methods available for tool execution.
- Last commit occurred over 450 days ago.
- Community-maintained repository lacking official vendor oversight.
Setup
{
"mcpServers": {
"ollama-mcp-bridge": {
"args": [
"-y",
"mcp-llm-bridge"
],
"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.
Compatibility
| Client | Local | Docker | Remote | Read-only |
|---|---|---|---|---|
| ChatGPT | ||||
| Claude Desktop | ||||
| Cursor | ||||
| VS Code | ||||
| Windsurf |
Frequently asked questions
›Which specific LLM model is required for this bridge?
The project is currently configured to use qwen2.5-coder:7b-instruct via Ollama.
›What environment variables are necessary for full tool functionality?
You must set BRAVE_API_KEY for search, GITHUB_PERSONAL_ACCESS_TOKEN for GitHub, and REPLICATE_API_TOKEN for Flux image generation.
›How do I authenticate the Gmail and Drive MCP servers?
You must execute the auth command using the index.js file path for the package, such as node path/to/gmail-drive/index.js auth.
›Where do I define the paths and settings for connected MCP servers?
All MCP server definitions, tool permissions, and LLM settings are configured within the bridge_config.json file.
Alternatives
Badge
Maintain this server? Add the live badge to your README.