← Back to AutoAdy

Run your Meta ads inside Claude

Your ad account, in the chat

AutoAdy ships a live MCP server that puts your Facebook and Instagram ad accounts inside Claude. Instead of tab-switching into Ads Manager, you ask questions and give instructions in plain language — Claude answers with live data from your account and, when you say so, makes the changes.

What you can do

  • Morning briefing — a full account audit on demand: spend, results, cost per result, and what needs attention, across every active campaign.
  • Diagnose a drop — when leads dry up or CPA spikes, the drop diagnostic tests competing theories against your data instead of guessing “creative fatigue”.
  • Pause the bleeders — get a kill list of ads burning budget without results, review it in the chat, and confirm before anything is paused.
  • Scale the winners — plan and apply budget increases that don’t reset Meta’s learning phase.

Connect on claude.ai

  1. Open Settings → Connectors → Add custom connector.
  2. Paste this URL:
https://www.autoady.io/api/mcp

Claude opens an AutoAdy sign-in window — approve access (OAuth, no API key needed) and you’re connected. Start a new conversation and ask about your ads.

Connect in Claude Code

Claude Code connects over HTTP with an API key in a header. Create a key at Settings → API Keys (autoady.io/settings), then run:

claude mcp add --transport http autoady https://www.autoady.io/api/mcp \
  --header "Authorization: Bearer YOUR_KEY"

Run /mcp to confirm autoady is connected, then ask “audit my ad account”.

Things to say

“Give me a morning briefing on my ad account — spend, results, and anything that needs attention.”

calls run_full_audit

“Why did my leads drop last week?”

calls diagnose_drop

“Show my top performing ads from the last 30 days.”

calls get_top_performing_ads

“Which ads should I pause? Let me review the list before anything happens.”

calls apply_all_kills

“Scale my best campaign by 20% without resetting learning.”

calls scale_budget_relative

“Which creative angles are winning right now — hooks, formats, audiences?”

calls analyze_creative_performance

Go deeper

Frequently asked questions

Do I need an API key to use AutoAdy in Claude?

Not on claude.ai — adding the connector opens an AutoAdy sign-in window and you approve access with OAuth. Claude Code and Claude Desktop use an API key instead: create one at autoady.io under Settings, in the API Keys tab.

Is it safe to let Claude change my campaigns?

Read tools (insights, audits, diagnostics) can't change anything and are available on every plan. Write tools require a Pro plan, are capped at 200 changes per hour per ad account, and destructive actions are annotated so Claude asks before acting. Kill lists always render for review first, and every change is logged in AutoAdy and reversible from AutoAdy or Ads Manager.

Which Claude surfaces does this work in?

claude.ai (custom connector with OAuth), Claude Code (HTTP MCP server with an API-key header), and Claude Desktop (one-click extension with interactive views — live ad cards, sortable tables, and a review-and-confirm pause flow).

What can Claude actually do with my ad account?

60+ tools: full account audits, performance breakdowns by campaign, ad set, and ad, drop diagnostics, creative analysis, budget changes, pausing and scaling, duplicating winners, and launching new ads — all against live data from your Meta account.

Does AutoAdy work with assistants other than Claude?

Yes. The same MCP server works with ChatGPT connectors in developer mode and any MCP-capable client — point it at https://www.autoady.io/api/mcp with your API key.

Privacy PolicyTerms of ServiceData DeletionContact