Brandlio MCP Documentation for Developers

Brandlio MCP is a remote Model Context Protocol (MCP) server that exposes 250+ tools across 20+ marketing platforms through a single, secure OAuth endpoint. Connect it once and any MCP-compatible AI — Claude, Cursor, ChatGPT, Gemini, and more — can query and act on your live marketing data.

Overview

Brandlio MCP connects your marketing and commerce stack to any AI assistant. It handles OAuth, token refresh, and platform rate limits, and returns clean, normalized data so the model answers with your real numbers instead of guessing.

There is nothing to install or self-host — it is a hosted remote MCP server. You point your AI client at one endpoint:

The MCP endpoint

https://mcp.brandlio.io/api/mcp

250+ tools across 20+ platforms, behind one connection.
OAuth 2.1 (PKCE) — no long-lived secrets in your client config.
Reads are free; write operations require confirmation and are audit-logged.
Tools follow the accounts you connect and respect your permissions.

Quick start

Create an account at brandlio.io and sign in.
Connect your platforms from the Brandlio dashboard (OAuth for most; API key for a few).
Add the connector URL to your AI client (see below).
Complete the browser sign-in when your client prompts you.
Start asking — the AI now answers with live data from your connected accounts.

Connect your AI client

Claude (web & desktop)

Settings → Connectors → Add custom connector → paste the endpoint URL → Connect, then complete the browser sign-in.

Claude Code

claude mcp add --transport http brandlio https://mcp.brandlio.io/api/mcp

Then run /mcp inside Claude Code and authenticate in the browser.

Cursor

Add Brandlio to ~/.cursor/mcp.json:

{
  "mcpServers": {
    "brandlio": {
      "url": "https://mcp.brandlio.io/api/mcp"
    }
  }
}

VS Code (GitHub Copilot)

Add Brandlio to .vscode/mcp.json:

{
  "servers": {
    "brandlio": {
      "type": "http",
      "url": "https://mcp.brandlio.io/api/mcp"
    }
  }
}

ChatGPT, Gemini, Windsurf & others

Add Brandlio as a custom/remote MCP server using the same endpoint URL. Any MCP-compatible client works.

Authentication

OAuth 2.1 (recommended)

PKCE-based browser sign-in. Your client never stores platform secrets — it holds only a short-lived token issued after you authenticate. This is the right choice for interactive clients like Claude, Cursor, and ChatGPT.

Personal Access Token (PAT)

For automation and headless use, generate a PAT in the dashboard and pass it as a Bearer header, or embed it in the URL path:

https://mcp.brandlio.io/api/mcp/t/<your-token>

PAT safety

PATs are brand-scoped, rate-limited, and audit-logged. Treat them like passwords — store them in a secret manager, never commit them, and rotate them if exposed.

Tools & capabilities

The server exposes 250+ tools, grouped by what they do:

Live reporting — campaign, keyword, creative, placement, geo, device, and hour-of-day performance.
Diagnostics — anomalies, wasted spend, and feed errors.
Campaign management — create and update campaigns, ad groups, ads, keywords, and audiences.
Optimization — negative keywords, bid and budget adjustments, and platform recommendations.
Commerce & revenue — orders, sales, and inventory from Shopify, WooCommerce, Amazon, and Stripe.
Analytics & SEO — GA4 and Search Console data.
Skills — pre-built playbooks (weekly reports, search-term negatives, RSA rewrites).
Workspace — Google Drive, Sheets, Calendar, Tag Manager, and BigQuery.

Supported platforms

Advertising

Google Ads
Microsoft Advertising
Meta
Amazon Ads
TikTok
LinkedIn
Pinterest
Snapchat
Reddit

Analytics & SEO

Google Analytics 4
Google Search Console

E-commerce

Shopify
WooCommerce
Amazon Seller Central
Google Merchant Center

Payments & workspace

Stripe
YouTube
Google Business Profile
Google Drive, Sheets & Calendar
Google Tag Manager
BigQuery

Skills (playbooks)

Skills are pre-built, multi-step workflows you can invoke in plain language. Call list_skills to see what is available; examples include:

Amazon weekly report
Google Ads negative-keyword optimization
Responsive search ad (RSA) headline rewrites
Search-term analysis
Account performance diagnostics

Utility tools

list_skills — browse the available playbooks.
get_credit_balance — check your remaining write-operation credits.
submit_feedback — send a note to support without leaving your AI client.

Example prompts

Which campaigns are above my $25 CAC threshold right now?
What's my Amazon ACOS this week vs last week?
Which Amazon search terms are getting clicks but no conversions?
Run the Amazon weekly report for Acme Coffee.
Compare my Search Console CTR this month vs last month.
What's my blended CAC across Meta and Google this week?

Security & privacy

OAuth 2.1 with PKCE eliminates long-lived secrets in client config.
Platform credentials are encrypted at rest.
Write operations require confirmation before they execute.
PAT usage is rate-limited and audit-logged.
You connect only the platforms you choose — nothing else is accessed.

Pricing

Reads are always free. You pay only for write operations, from a single shared pool across every connected platform. A free plan is available and paid plans include a free trial.

See current plans and limits on the pricing page.

FAQ

Do I need to self-host anything?

No. Brandlio MCP is a hosted remote server. You add one endpoint URL to your AI client and authenticate in the browser.

Is it safe to give an AI write access to my ad accounts?

Reads never change anything, and write operations require confirmation before they execute. OAuth 2.1 with PKCE means your client never holds long-lived secrets, and PAT usage is audit-logged.

Which AI clients are supported?

Any MCP-compatible client — Claude (web, desktop, Code), Cursor, VS Code with Copilot, ChatGPT, Gemini, Windsurf, and others — plus Brandlio's own agent.

How do I authenticate for automation?

Generate a Personal Access Token in the dashboard and pass it as a Bearer header or in the URL path (…/api/mcp/t/<your-token>). PATs are brand-scoped and rate-limited.

What does a write operation cost?

Writes draw from a shared monthly pool; reads are free. See the pricing page for current allowances by plan.

Support & links

Source & issues: github.com/brandlio-io/brandlio-mcp
Website: brandlio.io
Pricing: mcp.brandlio.io/pricing
In-app: call submit_feedback to reach support from your AI client.