Documentation

Build on TagMango
with AI assistance

TagMango MCP gives AI coding agents (Cursor, Claude Code, Claude Desktop, Lovable) deep knowledge of every TagMango API endpoint, authentication flow, and integration pattern — so they can write correct integration code on the first try.

Overview

What is this?

The TagMango MCP server is a Model Context Protocol server — a standard protocol that lets AI agents discover and query structured knowledge.

Instead of pasting API docs into every chat, you connect your AI tool once and it can browse all TagMango APIs, look up endpoint schemas, understand auth requirements, and get integration examples — all on demand.

🔌

Connect once

Add to Cursor, Claude Code, or Claude Desktop. No re-configuration needed.

📖

Full API registry

Every external and core-api endpoint with schemas, examples, and integration notes.

⚡

Always current

Registry is parsed from live OpenAPI specs and continuously enriched.

Setup

Connect your AI tool

Choose your tool and follow the setup steps below.

Add to ~/.cursor/mcp.json (create if it doesn't exist):

JSON 
{
  "mcpServers": {
    "tagmango-builders": {
      "url": "https://mcp.tagmango.com/mcp"
    }
  }
}

Restart Cursor after saving. The tools will appear automatically in Cursor Agent.

Run this command once (requires the Claude Code CLI):

Shell 
claude mcp add --scope user --transport http \
  tagmango-builders https://mcp.tagmango.com/mcp

Or add manually to ~/.claude/settings.json:

JSON 
{
  "mcpServers": {
    "tagmango-builders": {
      "url": "https://mcp.tagmango.com/mcp"
    }
  }
}

Add to ~/Library/Application Support/Claude/claude_desktop_config.json:

JSON 
{
  "mcpServers": {
    "tagmango-builders": {
      "url": "https://mcp.tagmango.com/mcp"
    }
  }
}

Restart Claude Desktop after saving.

Lovable supports remote MCP servers via Connectors. No config file needed.

STEP 1 Open Connectors

In your Lovable project, click Connectors in the left sidebar.

STEP 2 Search for "Custom MCP"

Use the search bar to find Custom MCP and select it.

STEP 3 Fill in the details

Enter a name (e.g. tagmango-builders), paste the server URL below, set authentication to No Authentication, and click Add & Authorize.

Server URL 
https://mcp.tagmango.com/mcp

Once authorized, Lovable's AI agent will have full access to the TagMango API registry.

Reference

Available tools

The MCP server exposes four tools. AI agents use these to discover and look up API information.

list_modules

Returns all API modules with descriptions, keywords, and use-case examples. Always call this first to discover which module covers a feature.

get_module_endpoints

Returns a summary list of all endpoints in a module — method, path, auth type, and context (frontend/backend).

get_endpoint_detail

Returns full details for 1–5 endpoint IDs: parameters, request/response schemas, integration notes, and usage examples.

submit_feedback

Lets AI agents report missing endpoints, inaccurate schemas, or unclear docs — and submit a session summary after completing a task.

Authentication

API authentication

TagMango has two API surfaces, each with different auth.

External API

Server-to-server. Uses a static API key as a Bearer token. Suitable for backend integrations where you control the server.

Authorization: Bearer <api-key>
x-whitelabel-host: yourdomain.com

Core API

User-facing SSO flow. TagMango passes a refresh token to your app URL. Exchange it for a short-lived access token.

POST /get-access-token
{ "refreshToken": "..." }
→ { "accessToken": "..." }

Full auth guides are available as MCP resources — your AI agent can read them with tagmango://guides/auth-external and tagmango://guides/auth-core-api.