Skip to main content

AtomicMCP overview

AtomicMCP lets supported AI clients understand and work with your Atomicat workspace. Instead of switching between dashboards, you can ask natural-language questions and request guided actions across your marketing stack.

Use AtomicMCP when you want to:

  • Ask for analytics summaries, comparisons, and optimization ideas.
  • Review pages, funnels, quiz flows, leads, sites, or videos.
  • Create or update supported assets through an AI client.
  • Clone or rebuild page ideas with guidance.
  • Keep a conversation going inside Claude, ChatGPT, or another supported MCP client.

How AtomicMCP fits with MIA

MIA is Atomicat’s built-in assistant. AtomicMCP is for external AI clients. Both help with Atomicat work, but they serve different habits:

  • Use MIA when you want help inside Atomicat.
  • Use AtomicMCP when you prefer to work from a supported AI client and bring Atomicat context there.

Supported workflow examples

AtomicMCP is most useful when the task has clear Atomicat context and a review step:

WorkflowExample request
Page review“Review my landing page for headline clarity, CTA placement, mobile readability, and missing trust elements.”
Page recreation“Use this reference URL as inspiration and create a new Atomicat page draft for my offer. Show the plan before making changes.”
Funnel analysis“Compare the last 30 days of funnel performance and identify the weakest step.”
Quiz planning“Turn this offer into a quiz funnel with qualifying questions, result logic, and a final CTA.”
Lead review“Summarize recent leads, common patterns, and any source or webhook issues I should check.”
Site/domain guidance“Review this site setup and tell me the next DNS or publishing step.”
Video performance“Compare my top videos this month and suggest which one to use in a campaign.”

For live assets, ask the AI client to plan first, then review changes in Atomicat before publishing, exporting, deleting, or sending traffic.

What you need

  • An Atomicat account with access to AtomicMCP.
  • A supported AI client, such as Claude or ChatGPT.
  • The MCP Server URL shown in your Atomicat account.
  • Permission to connect third-party tools for your workspace.
  1. Open AtomicMCP in Atomicat.
  2. Copy your MCP Server URL from the screen.
  3. Choose your AI client: Claude, ChatGPT, or desktop.
  4. Add Atomicat as a connector in that client.
  5. Sign in to Atomicat when prompted.
  6. Start with a simple prompt, such as asking for page or analytics context.

Review checkpoints

Before accepting an AtomicMCP-assisted change:

  • Confirm the target workspace, project, site, page, funnel, quiz, video, or lead list.
  • Ask for a short summary of the intended change.
  • Review copy, claims, tracking, forms, products, URLs, and mobile layout inside Atomicat.
  • Publish only after the normal Atomicat preview and test workflow passes.

Frequently asked questions

Is AtomicMCP an AI model?

No. AtomicMCP is a connection layer. Your AI client provides the conversation experience; AtomicMCP provides controlled access to supported Atomicat context and actions.

Does AtomicMCP replace the Atomicat dashboard?

No. Use AtomicMCP for assisted workflows and analysis. Use the Atomicat dashboard when you need full visual editing, billing, account management, or final review.

Which AI client should I use?

Claude is recommended for page creation, editing, cloning, and multi-step builder workflows. ChatGPT is also useful, especially for analytics, summaries, and general marketing analysis.