Skip to main content
Glama
Turbo.az MCP Server

Turbo.az MCP Server

by JavidGlyv

Turbo.az MCP Server

MCP (Model Context Protocol) server for Turbo.az automotive marketplace. This server enables Claude Desktop to search for cars and retrieve listing information from turbo.az.

⚠️ Important Note

Turbo.az blocks access from outside Azerbaijan. This server must run from an Azerbaijani IP:

  • Local computer (in Azerbaijan)

  • Via VPN with Azerbaijan IP

  • From a VPS located in Azerbaijan

Related MCP server: Begagnad MCP

🚀 Installation

1. Requirements

  • Python 3.10+

  • Google Chrome browser

  • pip

2. Server Setup

# Clone the repo or copy files
cd turbo-az-mcp

# Create virtual environment
python -m venv venv

# Activate
# Windows:
venv\Scripts\activate
# Linux/Mac:
source venv/bin/activate

# Install packages
pip install -e .

3. Test

# Run server manually
python -m src.server

4. Test MCP without LLM

Spawns the server and calls tools (requires Chrome):

uv run python scripts/test_mcp.py

🔧 Claude Desktop (local MCP, stdio)

Local-only: Claude Desktop runs the server as a subprocess. Do not use "Add custom connector" / Remote MCP URL.

  1. Find and open the config file:

    • Windows: %APPDATA%\Claude\claude_desktop_config.json

    • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json

    • Linux: ~/.config/Claude/claude_desktop_config.json

  2. Add to mcpServers section (or copy claude_desktop_config.example.json from the project and change cwd to your path):

With uv (recommended):

{
  "mcpServers": {
    "turbo-az": {
      "command": "uv",
      "args": ["run", "python", "-m", "src.server"],
      "cwd": "/home/javid/dev/turbo-mcp"
    }
  }
}

With Python / venv:

{
  "mcpServers": {
    "turbo-az": {
      "command": "python",
      "args": ["-m", "src.server"],
      "cwd": "/home/javid/dev/turbo-mcp"
    }
  }
}

Note: cwd is the full path to the project folder. Restart Claude Desktop.

Claude on Windows, Server in WSL

If Claude Desktop is on Windows but the project is in WSL, use wsl in config. WSL non-interactive shell may not have uv in PATH, so use venv python:

{
  "mcpServers": {
    "turbo-az": {
      "command": "wsl",
      "args": ["bash", "-c", "cd /home/javid/dev/turbo-mcp && .venv/bin/python -m src.server"]
    }
  }
}

Replace /home/javid/dev/turbo-mcp with your WSL project path. If venv is not .venv (e.g. venv), use venv/bin/python instead of .venv/bin/python.

Chrome in WSL: When server runs in WSL, Chrome/Chromium must be installed in WSL for Selenium (Windows Chrome won't work). In WSL terminal:

sudo apt update && sudo apt install -y chromium-browser

If chromium-browser requires snap, install Google Chrome for Linux or set CHROME_BINARY=/usr/bin/chromium (or installed path).

📋 Available Tools

1. search_cars

Search for cars.

Parameters:

  • make - Brand (BMW, Mercedes, Toyota, etc.)

  • model - Model (X5, E-Class, etc.)

  • price_min / price_max - Price range (AZN)

  • year_min / year_max - Year range

  • fuel_type - Fuel: benzin, dizel, qaz, elektrik, hibrid

  • transmission - avtomat, mexaniki

  • limit - Number of results (default: 10)

Example query: "Search for BMW X5 from 2020, price up to 50000 AZN on Turbo.az"

2. get_car_details

Detailed listing information.

Parameters:

  • listing_id - Listing ID or full URL

Example query: "Show details of this listing on Turbo.az: 12345678"

3. get_makes_models

List of makes and models.

Parameters:

  • make - Brand (to see its models, empty = all makes)

Example query: "What BMW models are available on turbo.az?"

4. get_trending

New/popular listings.

Parameters:

  • category - new, popular, vip

  • limit - Number of results

🐛 Troubleshooting

"403 Forbidden" error

  • Make sure you're running from Azerbaijan IP

  • Use VPN (Azerbaijan IP)

ChromeDriver error

  • Make sure Chrome browser is installed

  • webdriver-manager automatically downloads ChromeDriver

"DevToolsActivePort file doesn't exist" (WSL)

  • chromium-browser (snap) often doesn't work in WSL. Install Google Chrome:

    wget -q https://dl.google.com/linux/direct/google-chrome-stable_current_amd64.deb
    sudo dpkg -i google-chrome-stable_current_amd64.deb
    sudo apt install -f
  • Or with virtual display: sudo apt install xvfb then xvfb-run -a uv run python scripts/test_mcp.py

Timeout error

  • Check internet connection

  • Check if turbo.az is working

📄 License

MIT

Available Tools

4 tools
get_car_detailsC

Fetches detailed listing info from Turbo.az. Requires listing ID or URL.

Inputs: listing_id (required).

get_makes_modelsB

Fetches list of available makes and models on Turbo.az.

Inputs: make.

search_carsC

Car search on Turbo.az. Search by make, model, price range, year, etc.

Inputs: make, model, price_min, price_max, +5 more.

Tool Schema Changelog

Recent tool additions, removals, and schema changes observed during successful MCP inspections.

Loading tool schema changelog…

    TDQS

    B3.4/5.0

    Scored across 4 tools

    Disambiguation5/5

    Each tool has a clearly distinct purpose with no ambiguity: get_car_details retrieves specific listing details, get_makes_models provides metadata, get_trending shows popular listings, and search_cars enables filtered searches. The descriptions clearly differentiate their functions, making misselection unlikely.

    Naming Consistency5/5

    All tools follow a consistent verb_noun pattern with snake_case naming (get_car_details, get_makes_models, get_trending, search_cars). The verbs 'get' and 'search' are appropriately used and maintain a predictable structure throughout the set.

    Tool Count4/5

    With 4 tools, the server is well-scoped for a car listing domain, covering key operations like searching, viewing details, browsing trends, and accessing metadata. It feels slightly thin but reasonable, as each tool earns its place without redundancy.

    Completeness4/5

    The toolset covers core read-only operations for a car listing service: search, detail viewing, trend access, and metadata retrieval. Minor gaps exist, such as no tools for user interactions (e.g., saving favorites or contacting sellers), but agents can perform essential browsing tasks effectively.

    Related MCP Connectors

    Related MCP Servers

    • F
      license
      Not graded
      quality
      D
      maintenance
      Enables AI agents to interact with a retail cars database using SQL queries. Supports adding new car advertisements and searching existing cars based on user preferences through natural language.
      -
    • A
      license
      Not graded
      quality
      D
      maintenance
      Enables AI agents to search and retrieve listings from Sweden's largest second-hand marketplaces, Blocket and Tradera. Returns unified data including prices, images, seller information, and direct links to listings.
      9
      MIT
    • F
      license
      Not graded
      quality
      D
      maintenance
      Provides 25 interactive automotive intelligence tools for real-time market data, including VIN decoding, price predictions, and inventory analytics. It enables AI assistants to perform car searches, trade-in estimations, and market trend analysis using the Model Context Protocol.
      2
      -