GTMx MCP Server
Connect Claude Desktop, Cursor, or any MCP-compatible AI client directly to your GTMx brand data — brand context, Share of Voice, citations, and content drafts — without leaving your AI tool.
What is MCP?
Model Context Protocol (MCP) is an open standard that lets AI assistants connect to external data sources and tools. Instead of copy-pasting your brand data into every AI chat, MCP lets the AI query it directly.
The GTMx MCP server exposes your brand's intelligence — context, keywords, SOV data, citations — as structured tools that Claude, Cursor, or other AI clients can call in real time during a conversation.
Endpoint
You can verify this URL in Dashboard → Code → Functions → mcpServer
Authentication
Every request must include your GTMx session token as a Bearer token in the Authorization header.
Authorization: Bearer YOUR_SESSION_TOKENClient Setup
Claude Desktop
Edit ~/Library/Application Support/Claude/claude_desktop_config.json (macOS) or %APPDATA%\Claude\claude_desktop_config.json (Windows):
{
"mcpServers": {
"gtmx": {
"url": "https://gtmx.ai/api/functions/mcpServer",
"headers": {
"Authorization": "Bearer YOUR_SESSION_TOKEN"
}
}
}
}Cursor
Edit ~/.cursor/mcp.json:
{
"mcpServers": {
"gtmx": {
"url": "https://gtmx.ai/api/functions/mcpServer",
"headers": {
"Authorization": "Bearer YOUR_SESSION_TOKEN"
}
}
}
}Available Tools
Click any tool to see its parameters and an example call.
Typical workflow
Load brand context
Ask your AI client to call get_brand_context with your brand ID. It now knows your company description, GTM keywords, and audience personas.
Check your competitive position
Call get_sov_data and get_citations to pull your latest AI Share of Voice and citation data into the conversation.
Generate an article
Ask the AI to call generate_aeo_article with a target keyword. It uses your brand context automatically and deducts 1 credit.
Save it back to GTMx
Call create_content_draft with the generated content. It appears immediately in your AEO Content Drafts dashboard.
Brand-Scoped Public Endpoints
Every GTMx brand gets a public, no-auth MCP endpoint that any AI agent can call to learn about the brand. This makes your brand discoverable and callable in agentic workflows — Claude, GPT, Perplexity, or any custom AI agent can query your brand profile, articles, knowledge graph, and audience fit without needing credentials.
Quick discovery (GET)
A plain GET request returns a brand discovery card — no JSON-RPC needed. Useful for AI agents that crawl for MCP endpoints.
Claude Desktop / Cursor config
Any MCP client can connect to a brand endpoint with zero auth setup:
{
"mcpServers": {
"acme-corp": {
"url": "https://gtmx.ai/api/functions/brandMcpServer?brand=acme-corp"
}
}
}Available tools (all read-only, no credits)
Why this matters for AI visibility
When an AI agent needs to recommend a tool, compare vendors, or answer "what does X do," it looks for callable, structured data sources. A brand with a public MCP endpoint is directly callable — the agent gets authoritative, up-to-date answers from the brand itself rather than relying on scraped web content. This is the foundation of agentic brand distribution.
Security model
Authentication
Every request validates your Bearer token server-side via Base44 auth. No anonymous access.
Authorization
You can only access brands your account has permission to. Passing someone else's brand_id will return no data.
Credit protection
generate_aeo_article checks and deducts credits using the same logic as the in-app writer — plan limits and overrides both apply.
Rate limiting
generate_aeo_article is limited to 1 call per week per user, enforced server-side.
Write scope
Only create_content_draft writes data. No tool can delete brands, modify SOV data, or touch other users' records.