vertex-ai-mcp-server
Vertex AI MCP Server is a platform designed to facilitate interaction with Google Cloud's Vertex AI Gemini models. It offers a suite of tools for coding assistance and query answering, aiming to enhance development productivity with configurable and responsive AI capabilities.
Vertex AI MCP Server
This project is a Model Context Protocol (MCP) server focused on Google Cloud's Vertex AI Gemini models. It provides an array of tools for coding assistance and query answering, using both web search and internal model knowledge. The server features adjustable model settings, streaming APIs for better responsiveness, and basic API error handling. Users can conduct code analysis, recommend architecture patterns, perform dependency scans, and execute filesystem operations, among other tasks.
Features
- Access to Vertex AI Gemini models via MCP tools.
- Configurable model settings via environment variables.
- Streaming API and minimal safety filters.
Tools Provided
AI Tools
- Query answering with web search and direct model knowledge.
- Code analysis and security vulnerability scans.
Filesystem Operations
- File reading, writing, and editing.
Combined Operations
- Save generated project guidelines and query answers to files.
Prerequisites
- Node.js (v18+)
- Bun
- Google Cloud setup with enabled Vertex AI API.
Setup & Installation
- Place project files in desired location.
- Install dependencies with Bun.
- Configure environment variables as required.
- Build the server.
Usage
The server can be run directly using npx
or installed globally. Environment variables must be set appropriately.