Gemini Deep Research

MCP server and CLI tools for Google's Gemini Deep Research API (Interactions API).

About

The Deep Research API is an autonomous research agent powered by Gemini 3 Pro. It:

• Plans and executes multi-step research tasks
• Iteratively searches the web via Google Search
• Synthesizes findings into comprehensive reports with citations
• Typically takes 5-15 minutes per research task

📄 Example Output

See RESEARCH_SHOWCASE.md for a comprehensive demonstration of the tool's capabilities. This showcase was generated by using the Deep Research MCP Server to research the state of AI research agents—a meta-demonstration that includes:

• Parallel research workflows: Conducting web searches while waiting for Gemini's long-running task
• Gemini Deep Research output: An 8+ minute autonomous research task producing ~4,500 words with 43 sources
• Comparative analysis: Side-by-side comparison of different AI research approaches
• Real-world insights: Coverage of major players (Google, OpenAI, Anthropic, Perplexity), technical architectures (Agentic RAG, MCP, multi-agent systems), and future predictions

Setup

1. Install dependencies:

   pip install -r requirements.txt

2. Create a .env file with your API key:

   GEMINI_API_KEY=your_api_key_here
   

   Get your API key from Google AI Studio.

Usage

# Run with default query
python deep_research.py

# Run with custom query
python deep_research.py "What is quantum computing?"

The script polls the Interactions API endpoint until research completes.

API Details

• Endpoint: https://generativelanguage.googleapis.com/v1beta/interactions
• Agent: deep-research-pro-preview-12-2025
• Execution: Asynchronous (requires polling)

MCP Server

This repo also includes a remote MCP server for use with Claude Desktop.

Quick Start

1. Deploy the MCP server:

   cd mcp-server
   npm install
   npx wrangler login
   npm run deploy

2. Add to Claude Desktop as a custom connector:

   • URL: https://deep-research-mcp.foray-consulting.workers.dev/sse
   • Click Connect → Enter your Gemini API key in the form → Done!

See mcp-server/README.md for details.