Facebook Insights
→
Anthropic Claude
FACEBOOK INSIGHTS connector for Claude (MCP)
Ask Anthropic Claude about your Facebook Insights data using all available fields — without exporting a single CSV.
- +277 Facebook Insights fields plus calculated metrics — every dimension and breakdown Facebook Insights reports.
- Your Facebook Insights data already lives in Porter. Enable MCP and start asking Claude — no new integrations, no tokens, no servers.
- Cross your Facebook Insights performance with Google Analytics 4 revenue in one Claude conversation to see real outcomes, not just clicks.
Free forever plan · 14-day full trial · No credit card
How to connect Facebook Insights to Claude
-
Connect Facebook Insights to Porter
Log in to Porter with your Facebook Insights account and grant read-only access. Takes 2 minutes — no engineering, no exports.
-
Pick your ad account
Select one or blend multiple pages into a single data source — useful for agencies managing several clients.
-
Add Porter MCP to Claude
In Claude, paste the URL
https://mcp.portermetrics.com/mcpas a custom connector. No JSON-RPC, no local servers. -
Authorize Claude to read Porter
One click. Claude now has read-only access to your Facebook Insights data through Porter — same scope as your Porter account.
-
Ask in plain English
“Show me Facebook Insights performance trends over the last 14 days.” Claude returns a filtered table in seconds.
+277 Facebook Insights fields, ready for Anthropic Claude
Every field in Facebook Insights Manager — plus calculated metrics, every attribution window, and creative preview URLs — accessible to Claude on demand.
All Ads Manager fields (and more)
277 fields total — every native Facebook Insights field plus calculated metrics that aren’t offered natively.
Every breakdown
Segment by campaign, ad set, creative, audience, placement, device, region, hour, date — every breakdown Facebook Insights supports.
Creative previews and assets
Ad preview URLs plus links to creative assets — Claude can reason over performance + creative in the same answer.
Marketer-friendly naming
Fields named exactly as in Facebook Insights (e.g. “Cost” not cost_micros). Claude understands them natively.
Full granularity
Hour, day, week, month, year — plus every dimension Facebook Insights reports.
Cross-channel blending
Same field naming standard across Porter’s connectors — Claude can join Facebook Insights with GA4, BigQuery, or other sources in one prompt.
- Page likes
- Page likes age
- Page likes gender
- Page followers age
- Page followers gender
- Page total followers
- Post engagements
- Likes churn rate
- Likes growth rate
- Net likes growth
- NonViralReachPerc
- OrganicReachPerc
- Post actions: Anger
- Post actions: Haha
- Post actions: Like
- Post actions: Love
- Post actions: Sad
- Post actions: Wow
- Reel backdated time
- Reel backdated time granularity
- Reel blue reels play count
- Reel content category
- Reel creation time
- Reel views
- Video backdated time
- Video backdated time granularity
- Video content category
- Video created time of day
- Video description
- Video embed html
- City
- Consumption type
- Country code
- Country icon
- Country icon url
- Country name
- Month
- Week
- Year
- Month and year
- Week and year
- Date
- Review created date
- Has review rating
- Has review
- Recommendation type
- Review rating
- Review text
- Post description
- Post Image URL
- Shared post link
- Post name
- Post Type String
Copy-paste prompts for Facebook Insights × Claude
Add Porter MCP to Claude (paste https://mcp.portermetrics.com/mcp), open a new conversation, and paste any prompt below.
Organic vs. Paid Reach Attribution Diagnosis
Compare my Organic reach and Paid reach for the last 90 days. Flag any week where Organic reach dropped more than 15% while Paid reach stayed flat or grew. For each flagged week, show me the matching Viral Reach and Total reach so I can tell if the drop was algorithmic, budget-driven, or viral decay. Output a table with week, organic reach, paid reach, viral reach, and a one-sentence diagnosis per row.Weekly Executive Summary for Leadership
Build my weekly executive summary for the last 7 days. Start with Page total reach, then break it into Organic Impressions, Paid impressions, and Viral Impressions. Add Page Engaged Users, Page new followers, and Page new unfollows. Compare each metric to the prior week in percent-change terms. Finish with a 3-bullet "so what" section that tells me whether the page is growing, stable, or at risk. Output as a clean markdown summary I can paste into an email.Video Hook Drop-off Audit
Audit my video performance for the last 30 days. For each video, show Page video views, Page video views 10s, and Page video complete views 30s. Calculate the drop-off between each stage and flag the video with the biggest gap between Page video views and Page video views 10s. Also compare Page video autoplayed views vs. Page video clicked to play views for that video. Give me 2 specific hook-rewriting suggestions based on the drop-off pattern. Output as a ranked list.Sentiment Swing Anomaly Detector
Chart my weekly Page actions: Anger against Page actions: Love for the last 60 days. Calculate a sentiment ratio (love / anger) per week and flag the week with the biggest swing — either a 25% drop in the ratio or any week where anger exceeds love. For that flagged week, pull Page new unfollows (Unique) and Page negative actions unique to see if the mood shift translated into churn. Output a table with week, love count, anger count, ratio, unfollows, and a one-line alert status.Creative Fatigue Early Warning
For the last 45 days, group my posts by Post type and calculate average Post engagements, Post shares, and Post comments count per type. Compare the last 14 days against the prior 14 days for each type. Flag any Post type whose Post engagement rate dropped more than 20% week-over-week for two consecutive weeks. For each flagged type, list the 3 most recent Post message examples and suggest 2 format refreshes. Output as a ranked table with trend arrows.Audience Demographics Overlap Check
Compare my Page likes age and Page likes gender breakdowns against Page followers age and Page followers gender for the last 30 days. Highlight any age bracket or gender where followers grew but likes stayed flat — that signals a reach gap. Cross-reference with Page fans online to see if timing explains the mismatch. Output a side-by-side comparison table plus a 2-sentence recommendation on which demographic to target or adjust posting times for.Reels vs. Legacy Video Cross-Format Benchmark
For the last 45 days, pull Reel blue reels play count and Reel post video avg time watched for every reel. Pull Page video views and Post average time video viewed for every regular video from the same period. Normalize both sets by dividing watch time by video length to get an attention-efficiency score. Recommend whether I should shift production budget toward reels or stick with long-form video based on which format has the higher efficiency score and lower decay rate. Output as a summary table plus a one-paragraph recommendation.Multi-Account Agency Roll-up Dashboard
Roll up the last 30 days across all Facebook pages I have connected. For each Page name, show Page total impressions, Organic reach, Paid reach, Page Engagements, Page total followers, Page new followers, and Page new unfollows. Calculate an engagement rate (Page Engagements / Page total impressions) and a net growth rate (Page new followers - Page new unfollows) / Page total followers) per page. Rank pages by net growth rate and flag any page where Paid reach exceeds Organic reach by more than 2:1. Output as a master table sorted by net growth.Video Ad Break Earnings & Budget Pacing Tracker
For the last 30 days, match my top 10 videos by Page video views with their Post video ad break earnings, Post video ad break ad impressions, and Post video ad break ad cpm. Calculate earnings per 1,000 views for each video. Flag any video with high views but low earnings — that means the ad fill or CPM is underperforming. Cross-reference with Post video views time to see if watch duration explains the CPM gap. Output a ranked table with a "pacing status" column: on-track, under-earning, or viral-but-unmonetized.Review-Driven Reputation Alert System
Scan the last 30 days for every review where Recommendation type is negative or Review rating is 1–2 stars. Pull Review text, Review created date, and Reviewer name. For each negative review date, check Page messages new conversations unique and Page new unfollows (Unique) on that same day and the following 2 days. If either metric spikes more than 50% above the 7-day average, flag it as a reputation incident. Output an alert table with review date, rating, review excerpt, conversation spike %, unfollow spike %, and a red/yellow/green severity badge.Facebook Insights MCP safety
Porter authenticates to Facebook Insights with the read-only scope you already approved for your reports. Anthropic Claude queries Porter — not Facebook Insights directly — so every conversation is rate-limited and cached at Porter side. No new permissions, no ban risk.
- Caching. Porter caches recent query results for 15 minutes per (account, field set) so repeated Claude prompts don’t hit the API again.
- Batching. Multiple field requests are merged into a single API call when possible (max 50 per call).
- Backoff & retry. Exponential backoff on 429 responses (1s → 2s → 4s → 8s).
- Per-account quotas. Porter tracks per-account daily usage and pauses queries before hitting 80% of the 200 calls/hour limit.
- Read-only enforcement. Porter never requests write scopes from Facebook Insights, so even compromised tokens cannot modify campaigns, budgets, or ad copy.
- Audited tools. The MCP exposes 6 read-only methods:
get_blend_id,list_accounts,list_categories,list_data_sources,list_fields,query_data.
GDPR compliant · 1,600+ marketers trust Porter with their Facebook Insights data.
Start free. Pay per data source account
- +277 Facebook Insights fieldsEvery native field plus calculated metrics — accessible to Claude in plain English.
- All attribution windowsCompare 1d-click, 7d-click, 28d-click and 1d_click+1d_view side by side from one connection.
- Same plan, every destinationSend to Anthropic Claude, ChatGPT, Looker Studio, Google Sheets, BigQuery and Slack — one subscription.
- Unlimited usersNo extra cost per team member or seat.
Number of Facebook Insights pages
Billed annually · $12.5/account
Unlimited 14-day free trial + Free forever plan
Other connectors & destinations
Porter supports 15+ data sources and 10+ destinations. Mix and match in one subscription — no extra seats or hidden per-destination fees.
Questions about Facebook Insights + Anthropic Claude
What is MCP and how does it connect Facebook Insights to Claude?
Do I need technical skills or coding to set this up?
Will connecting my Facebook Insights account get it banned?
What Facebook Insights metrics can I analyze in Claude?
Can I create or edit campaigns from Claude, or is this read-only?
I already use Porter for Facebook Insights reporting — do I need to reconnect?
How much does the Facebook Insights MCP integration cost?
Chat with your Facebook Insights data in Claude
Your Facebook Insights data already lives in Porter. Enable MCP and start asking Claude about campaigns, creatives, and ROAS — no new integrations, no tokens, no servers. Get live, conversational access to all 277 Facebook Insights fields including creative previews, every attribution window, and multi-account blends.
Free forever plan · 14-day full trial · No credit card