M cenaly.ru
🔗 Integrations

🔌 Connect AI Tools to MCP

Step-by-step guides for Claude, ChatGPT, Cursor, VS Code, Windsurf, Cline and more

Documentation

Connecting AI Tools to cenaly.ru MCP Server

Step-by-step guides for connecting popular AI assistants to your restaurant data via the MCP protocol.

Prerequisites#

Before connecting any tool, you need:

  • Server URL: https://api.cenaly.ru/mcp
  • API Key: generated in the admin panel at admin.cenaly.ru → ⚙️ Settings → 🔑 Access → 🤖 MCP API Keys

💡 See the MCP Server User Guide for detailed instructions on generating your API key.

⚠️ There is no MCP server in the Russian contour. On cenaly.ru it is not deployed (/mcp/health replies 404 not found), so the "MCP API Keys" section is not shown in Settings there — there is nothing to connect. In every other contour the address shown in the panel works; if your account lives outside the main contour, you may not see your own data there yet — splitting the storage per contour is in progress.

Claude Desktop#

The desktop app by Anthropic.

Step 1. Open Claude Desktop

Step 2. Go to Settings (⚙️) → Developer → Edit Config

Step 3. The file claude_desktop_config.json will open. Add:

{
  "mcpServers": {
    "meni": {
      "url": "https://api.cenaly.ru/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Step 4. Save the file and restart Claude Desktop

Step 5. In a new chat, a 🔨 (tools) icon will appear at the bottom — click it and verify the meni MCP server is connected

📁 Config file location:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json
  • Linux: ~/.config/Claude/claude_desktop_config.json

Claude Code (CLI)#

Anthropic's terminal-based AI assistant.

Option 1 — Via command (recommended):

claude mcp add meni \
  --transport http \
  --url https://api.cenaly.ru/mcp \
  --header "Authorization: Bearer YOUR_API_KEY"

Option 2 — Config file:

Edit file ~/.claude/settings.json:

{
  "mcpServers": {
    "meni": {
      "url": "https://api.cenaly.ru/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Verify: Run claude and ask "What are my locations?" — Claude will call my_locations.

Cursor#

AI code editor with built-in MCP support.

Option 1 — Via UI:

  1. Open Cursor
  2. Go to Settings (Cmd/Ctrl + ,) → MCP
  3. Click "+ Add new MCP Server"
  4. Fill in:
    • Name: meni
    • Type: HTTP
    • URL: https://api.cenaly.ru/mcp
  5. Click Save
  6. Open the .cursor/mcp.json file and add the authorization header (see below)

Option 2 — Config file:

Create .cursor/mcp.json in your project root:

{
  "mcpServers": {
    "meni": {
      "url": "https://api.cenaly.ru/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Verify: In Cursor chat (Agent mode), ask "Show my locations on cenaly.ru" — Cursor will connect to the server.

💡 Cursor supports MCP in Agent mode. Make sure Agent mode is selected, not Ask.

VS Code (GitHub Copilot)#

The GitHub Copilot extension for VS Code supports MCP servers.

Option 1 — Project file (recommended):

Create .vscode/mcp.json in your project root:

{
  "servers": {
    "meni": {
      "type": "http",
      "url": "https://api.cenaly.ru/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Option 2 — Global settings:

Open Settings (Cmd/Ctrl + ,) → search for mcp → open settings.json and add:

{
  "mcp": {
    "servers": {
      "meni": {
        "type": "http",
        "url": "https://api.cenaly.ru/mcp",
        "headers": {
          "Authorization": "Bearer YOUR_API_KEY"
        }
      }
    }
  }
}

Verify: Open Copilot Chat (Ctrl+Alt+I), select Agent mode, and ask: "What are my locations on cenaly.ru?"

💡 MCP servers only work in Agent mode (not in Edits or Ask).

Windsurf#

AI editor by Codeium with MCP support.

Step 1. Open Windsurf

Step 2. Click the 🔨 (hammer) icon in the Cascade panel, then Configure

Or navigate to: Settings → Cascade → MCP Servers

Step 3. The mcp_config.json file will open. Add:

{
  "mcpServers": {
    "meni": {
      "serverUrl": "https://api.cenaly.ru/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Step 4. Save and restart Windsurf

Verify: In the Cascade chat, the 🔨 icon will show available MCP tools.

📁 Config file location: ~/.codeium/windsurf/mcp_config.json

Cline (VS Code Extension)#

Autonomous AI agent for VS Code.

Step 1. Open Cline in VS Code (side panel)

Step 2. Click MCP Servers (🔌 icon) at the top of the Cline panel

Step 3. Click "Remote Servers", then "Add Remote MCP Server"

Step 4. Fill in:

  • Server URL: https://api.cenaly.ru/mcp
  • Name: meni
  • Headers (JSON format):
{
  "Authorization": "Bearer YOUR_API_KEY"
}

Step 5. Click Save

Or edit the cline_mcp_settings.json file directly:

{
  "mcpServers": {
    "meni": {
      "url": "https://api.cenaly.ru/mcp",
      "headers": {
        "Authorization": "Bearer YOUR_API_KEY"
      }
    }
  }
}

Verify: A green indicator next to the server means successful connection.

Continue (VS Code / JetBrains)#

Open-source AI assistant for VS Code and JetBrains IDEs.

Step 1. Open Continue settings: click ⚙️ in the Continue panel

Step 2. Switch to the MCP tab

Step 3. Click "Add MCP Server" and select type Streamable HTTP

Step 4. Or edit the file ~/.continue/config.yaml:

mcpServers:
  - name: meni
    url: https://api.cenaly.ru/mcp
    headers:
      Authorization: "Bearer YOUR_API_KEY"

Verify: In Continue chat, ask "Show my locations" — Continue will use MCP tools.

ChatGPT (Connectors)#

OpenAI ChatGPT supports external MCP servers via Connectors.

Step 1. Open chatgpt.com

Step 2. Go to Settings → Connectors → Add Connector

Step 3. Select type MCP

Step 4. Fill in:

  • URL: https://api.cenaly.ru/mcp
  • Authentication: Bearer Token
  • Token: YOUR_API_KEY

Step 5. Click Save

Verify: In a new ChatGPT chat, ask "What are my locations on cenaly.ru?" — ChatGPT will use MCP tools.

⚠️ Connectors are available for ChatGPT Plus / Team / Enterprise subscribers.

Troubleshooting#

Test Your Key#

You can verify your API key works before configuring any tool:

curl https://api.cenaly.ru/mcp/health \
  -H "Authorization: Bearer YOUR_API_KEY"

Successful response: {"status":"ok"}

Common Issues#

Problem Solution
Connection refused Verify the URL is exactly https://api.cenaly.ru/mcp
404 not found instead of {"status":"ok"} The server is not running in your contour — see the warning at the top of this page
401 Unauthorized Check your API key — it may be revoked or mistyped
Tools not showing Restart the AI tool after saving the config
Key doesn't work Ensure format is Bearer <key> (with space after Bearer)
Key lost Revoke it in admin panel and generate a new one

General Tips#

  • Copy the API key without spaces or extra characters
  • Header format is always: Authorization: Bearer your_key
  • The panel (admin.cenaly.ru → Settings → Access) shows which keys exist, their names and creation dates; the key itself and its activity are not shown there — when in doubt, revoke it and generate a new one
  • You can have up to 10 keys — create separate keys for each tool
  • Revoked keys stop working immediately
Documentation
Guest demo menu Admin panel demo Start free