Meta Ads MCP: Connect Claude and ChatGPT

Ad network MCP guide

Connect Meta Ads to Claude or ChatGPT with Lemon AI’s unified MCP and API. Analyze creative performance without mistaking CTR changes for fatigue.

You can analyze Facebook and Instagram advertising in Claude or ChatGPT by connecting Meta reporting through MCP. With Lemon AI, connect Meta once and authorize the same MCP server you use for your other ad networks. The assistant can retrieve creative performance, inspect assigned attributes, and investigate an apparent winner without another round of copying Ads Manager tables into chat.

The connection solves data retrieval. The analysis still needs the right account, dates, report grain, and attribution definition. A confident answer to the wrong report is not a useful creative recommendation.

Compare the three ways to bring Meta data into chat

The Meta Marketing API Insights interface is the underlying reporting reference for a direct integration. When evaluating a Meta Ads MCP server, establish who operates it, which account it reads, and whether its tools can modify campaigns. A connector that can create ads has a different purpose from a read-only reporting connection.

For a direct connector, authorize the intended Meta business and ad account, inspect its discovered tools, and retrieve a small report before asking for recommendations. Use the connector's current official instructions for setup and eligibility. Do not assume two servers called “Meta Ads MCP” expose the same data or permissions.

Connect Meta Ads to Lemon

1. Open Connections in Lemon and choose Meta Ads.
2. Complete the Facebook authorization with a user who has access to the relevant business and ad accounts. Review the requested permissions in that flow.
3. Select the accounts to sync. After the initial sync, confirm a known campaign and date range in Creative Analytics.
4. Open API & MCP and connect Claude or ChatGPT using the shared server URL.

Lemon's MCP tools are read-only. That describes what the assistant can do through this interface; it does not imply that Meta's account-connection authorization requests only a single read permission. Keep those two permission boundaries distinct.

Add Lemon to Claude or ChatGPT

Use the same remote MCP server for all the networks you connect in Lemon:

https://api.lemon-ai.com/mcp

In Claude, open Customize → Connectors → + Add → Add custom connector. Enter a name and the URL, continue, and sign in to Lemon. If Claude asks how to register the OAuth client, choose Register automatically. Enable the connector in the conversation. Organization-managed accounts may need an owner to add it first. See the Claude setup instructions.

In ChatGPT, open Plugins, choose Add custom MCP server, and enter the URL. Select OAuth authentication, authorize Lemon, and install the resulting personal plugin. Mention it in a Work chat to use its tools. Availability depends on your workspace settings; follow the current ChatGPT MCP instructions if the interface differs.

Start by asking the assistant to call list_ad_accounts and let you select the intended account. Then use list_filter_values to discover dates and filters. Your ad-network sign-in connects data to Lemon; this second authorization lets the AI app read it. Do not paste network credentials into chat.

Start with the creative, not an account-wide verdict

Use Lemon AI to discover my Meta accounts and available campaigns. After I select an account, compare video assets for the same app, country, and two complete weeks. Return impressions, clicks, cost, and CTR. Show the scope and any missing or truncated data. Do not change campaigns.

This uses list_ad_accounts, list_filter_values, and query_creative_performance. Review the account and exact filter values the assistant selects. A campaign name containing “US” is not evidence that its delivery was restricted to the US.

Move to list_creative_attributes and attribute_insights only after the population is clear. Lemon's creative labels give the assistant something more useful than an unstructured list of ad names: it can investigate hook, format, subject, or another attribute your account actually defines.

Keep reporting time attached to the outcome

Meta results depend on what was attributed and when the report counts it. Lemon's current Meta reporting does not expose the old impression-time click_1d, click_7d, and click_28d cohort windows as available outcomes. Do not instruct the agent to query those suffixes simply because an old example contains them.

Ordinary Meta conversion reporting and a post-install cohort are different concepts. Ask discovery what is available through the interface you are using. If a required value is unavailable, the right answer is to say so and use the matching source report, not relabel a different column.

Synthetic example, not customer data: a creative delivered 20,000 impressions and 400 clicks in one period, followed by 20,000 impressions and 300 clicks in the next. CTR moved from 2.0% to 1.5%, a decline of 0.5 percentage points, or 25% relative to the starting rate.

That arithmetic supports “click response fell in this scope.” It does not establish that purchases fell, that ROAS fell, or that fatigue caused the change. Those claims need their own evidence. See the creative-fatigue measurement protocol.

Ask why performance changed

For this Meta account and the same dates, inspect the available hook labels. Compare CTR with attribute_insights, preserving its evidence filters. Show coverage and examples of the creatives behind each finding. Do not treat a label-performance association as causal.

For the creative with falling CTR, compare the same campaign and country in both periods. Tell me whether the apparent decline survives the matched scope. If the data cannot separate fatigue from delivery changes, say what additional comparison is needed.

These prompts produce a useful next investigation even when the original “winner” disappears. A missing finding is not a reason to manufacture a recommendation.

Use the same reporting layer through the API

For a script or another AI app, create a Lemon API key in API & MCP. Retrieve your account IDs with GET /v1/ad-accounts, then make a scoped read. This example uses a placeholder account and dates; substitute values returned for your account.

curl --get 'https://api.lemon-ai.com/v1/creatives/performance' \
  --header "Authorization: Bearer $LEMON_API_KEY" \
  --data-urlencode 'ad_network_account_id=YOUR_ACCOUNT_ID' \
  --data-urlencode 'date_from=2026-10-01' \
  --data-urlencode 'date_to=2026-10-07' \
  --data-urlencode 'metric=impressions' \
  --data-urlencode 'metric=clicks' \
  --data-urlencode 'metric=cost' \
  --data-urlencode 'sort_by=cost' \
  --data-urlencode 'sort_order=descending' \
  --data-urlencode 'limit=25'

MCP and REST expose Lemon's reporting definitions. Inspect missing fields, result limits, and any outcome metadata before interpreting a result. A limited top-creatives response is not an account total. Use the API documentation for supported fields and pagination.

Troubleshoot account, media, and metric gaps separately

An absent account is an authorization or selection question. A missing recent period is a sync or source-availability question. An unavailable outcome is a metric-capability question. Reconnecting an account will not turn an unsupported metric into a valid one.

A Meta asset can have performance without an accessible playable video URL. A thumbnail does not prove that the assistant viewed the video. Ask the assistant to distinguish stored attributes from any new visual interpretation and to identify inaccessible media.

For recurring analysis, connect your ad accounts. The MCP and API guide explains how the same connection works across networks, and Reports shows the product surface behind it.