gemini-mcp-server

Gemini MCP Server

Multi-context processing (MCP) server using Gemini API for analyzing webpages.

Features

• Screenshot and analyze webpages
• Tag and metadata for screenshots
• Retrieve screenshots and associated data via API

API Endpoints

Process a webpage

POST /process

Request body:

{
  "url": "https://example.com",
  "prompt": "Describe what you see on this webpage"
}

Get screenshots

Get screenshot by ID

GET /screenshots/:id

Example:

GET /screenshots/123

Response:

{
  "id": 123,
  "url": "https://example.com",
  "prompt": "Describe what you see on this webpage",
  "created_at": "2023-06-15T12:34:56Z",
  "tags": ["initial_load", "page_entry"],
  "metadata": {
    "pageState": "initial",
    "pageTitle": "Example Domain"
  },
  "download_url": "https://yoursupabaseproject.supabase.co/storage/v1/object/public/screenshots/screenshot_1234.png"
}

Get screenshots with filters

GET /screenshots?tag=initial_load

or

GET /screenshots?key=pageTitle&value=Example%20Domain

Response:

{
  "count": 2,
  "screenshots": [
    {
      "id": 123,
      "url": "https://example.com",
      "prompt": "Describe what you see on this webpage",
      "created_at": "2023-06-15T12:34:56Z",
      "tags": ["initial_load", "page_entry"],
      "metadata": {
        "pageState": "initial",
        "pageTitle": "Example Domain"
      },
      "download_url": "https://yoursupabaseproject.supabase.co/storage/v1/object/public/screenshots/screenshot_1234.png"
    },
    // Additional screenshots...
  ]
}

Installation

1. Clone the repository
2. Create a .env file based on .env.example
3. Run npm install
4. Run npm run build
5. Start the server with npm start