Skip to content

Synter-Media-AI/cross-platform-ads-agent

Repository files navigation

Cross-Platform Ads MCP Starter Kit — Manage All Ad Platforms with AI

MCP Compatible License: MIT Platform: Multi-Platform

One agent, every ad platform. Open this repo in Amp, Cursor, or VS Code and manage Google, Meta, LinkedIn, Microsoft, Reddit, TikTok, X, Amazon, Spotify, and The Trade Desk — all from a single chat window with unified reporting and cross-channel budget optimization.


Why Cross-Platform Management?

Most businesses advertise on 3-5 platforms. Each platform has its own dashboard, its own metrics definitions, its own attribution model, and its own API. Comparing performance across platforms means opening 5 tabs, exporting 5 CSVs, normalizing the data in a spreadsheet, and spending 2 hours every Monday on reporting.

Meanwhile, budget allocation decisions that could save thousands — like shifting $2,000 from underperforming LinkedIn to high-ROAS Meta — take days instead of minutes because the data lives in silos.

A cross-platform AI agent collapses all these silos. One question — "Which platform has the best ROAS this month?" — and you get a normalized comparison across every channel.

Best for: Agencies managing multi-platform campaigns, marketing teams running 3+ ad channels, anyone who needs a unified view of advertising performance.


Quick Start (30 Seconds)

Amp / Cursor / VS Code (Copilot)

  1. Get a free API key at syntermedia.ai/developer
  2. Set the key:
    export SYNTER_API_KEY=syn_your_key_here
  3. Open this repo in your editor
  4. Start chatting — MCP tools are pre-configured in .mcp.json

Claude Desktop

Copy claude_desktop_config.json to your Claude config directory and replace the API key:

  • macOS: ~/Library/Application Support/Claude/claude_desktop_config.json
  • Windows: %APPDATA%\Claude\claude_desktop_config.json

What You Can Do

Capability Example
Unified Reporting "Show me spend, ROAS, and CPA across all platforms for the last 30 days"
Budget Reallocation "Move $2,000 from LinkedIn to Meta — Meta is outperforming 3x"
Cross-Platform Comparison "Which platform has the lowest CPA for lead generation?"
Weekly Summary "Generate my Monday morning report across all channels"
Launch Campaigns "Launch the same promotion on Google, Meta, and LinkedIn simultaneously"
Pause Underperformers "Pause any campaign with ROAS below 1.5x on any platform"
Channel Attribution "How do customers interact across platforms before converting?"
Anomaly Detection "Did anything unusual happen across my campaigns this week?"
Budget Pacing "Am I on track to spend my full monthly budget across all platforms?"
Competitive Intelligence "Compare our share of voice across Google and Meta"

Example Conversations

1. Monday Morning Cross-Platform Report

You: Give me the weekly report.

Agent: Here's your cross-platform performance for the last 7 days:

Overview:

Platform Spend Revenue ROAS CPA Conv.
Google Ads $4,200 $18,900 4.5x $28 150
Meta Ads $3,800 $15,200 4.0x $22 173
LinkedIn Ads $2,500 $4,250 1.7x $89 28
Microsoft Ads $800 $4,400 5.5x $19 42
TikTok Ads $1,200 $3,600 3.0x $15 80
Total $12,500 $46,350 3.7x $26 473

Week-over-week changes:

  • 📈 Microsoft Ads: ROAS up 38% (new campaign hitting stride)
  • 📈 TikTok: CPA down 22% (new creative winning)
  • 📉 LinkedIn: ROAS down 15% (investigate creative fatigue)
  • ➡️ Google & Meta: Stable

Action items:

  1. Increase Microsoft budget. 5.5x ROAS is your best performer but only 6% of spend. It can absorb $1,500-2,000 more.
  2. Investigate LinkedIn. ROAS dropped from 2.0x to 1.7x. Likely creative fatigue — refresh creative or narrow audience.
  3. Scale TikTok. CPA is the lowest ($15) and dropping. Increase budget by 20%.

2. Data-Driven Budget Reallocation

You: I have $15,000/month total. How should I split it?

Agent: Based on the last 60 days of performance data across all your platforms:

Current allocation vs Recommended:

Platform Current Current ROAS Recommended Expected ROAS
Google Ads $5,000 (33%) 4.2x $5,500 (37%) 4.0x
Meta Ads $4,500 (30%) 3.8x $4,000 (27%) 3.9x
LinkedIn $3,000 (20%) 1.7x $1,500 (10%) 2.0x
Microsoft $1,000 (7%) 5.5x $2,000 (13%) 4.8x
TikTok $1,500 (10%) 3.0x $2,000 (13%) 2.8x

Changes explained:

  • Microsoft +$1,000: Your best ROAS (5.5x) is severely underfunded. Doubling the budget will still maintain strong returns.
  • LinkedIn -$1,500: At 1.7x ROAS, LinkedIn is barely profitable. Cut to just your best-performing ABM campaigns.
  • TikTok +$500: Lowest CPA and improving trends. Worth testing more budget.
  • Google +$500: Stable performer, can absorb incremental budget.
  • Meta -$500: Slight trim — reallocate to higher-ROAS channels.

Expected total impact: Blended ROAS moves from 3.5x to 3.9x. That's an additional $6,000 in monthly revenue for the same $15,000 spend.

3. Same Campaign, Multiple Platforms

You: We're launching a new product next week. I want ads on Google, Meta, and TikTok simultaneously.

Agent: I'll create a coordinated launch across all three platforms. Here's the plan:

Google Ads:

  • Search campaign: "new [product category]" keywords
  • PMax with product feed for Shopping
  • Budget: $40/day

Meta Ads:

  • Advantage+ campaign with product images and lifestyle shots
  • Target: Lookalike audience from existing customers
  • Budget: $35/day

TikTok:

  • In-Feed video ads (UGC style)
  • Target: Interest-based + broad (let algorithm optimize)
  • Budget: $25/day

Coordination rules:

  • All campaigns launch at midnight PT on launch day
  • Consistent messaging across platforms (adapted for format)
  • Shared UTM parameters for cross-platform attribution in GA4
  • Daily check-ins for the first week: kill losers early, scale winners

Week 1 review cadence:

  • Day 2: First data check (enough for directional signals)
  • Day 4: Kill any campaign with CPA 3x above target
  • Day 7: Full analysis — which platform won?

Shall I create all three campaigns now?

4. Anomaly Detection Across Platforms

You: Anything weird happen with my campaigns this week?

Agent: Scanning all platforms for anomalies in the last 7 days...

🚨 Anomalies detected:

1. Google Ads — Spend spike on Wednesday

  • Campaign "Brand — Exact" spent $420 (normal: $80-120/day)
  • Cause: A competitor started bidding on your brand name, driving up CPCs by 3.5x
  • Action needed: Add competitor as negative keyword OR accept higher brand CPCs

2. Meta Ads — CTR drop on Friday

  • Your top ad set CTR dropped from 2.1% to 0.7% overnight
  • Cause: Creative frequency hit 5.2 — classic creative fatigue
  • Action needed: Introduce 2-3 new creatives immediately

3. TikTok — 0 conversions Saturday-Sunday

  • Platform shows clicks and views, but 0 purchases
  • Cause: Your TikTok Pixel stopped firing on Saturday at 2pm
  • Action needed: Check if a website deploy broke the pixel. Conversions likely happened but weren't tracked.

No anomalies on: LinkedIn, Microsoft Ads (both performing within normal range).

Want me to fix the TikTok Pixel and refresh the Meta creative?

5. Cross-Platform Attribution Analysis

You: Are customers touching multiple platforms before converting?

Agent: Let me analyze your GA4 multi-channel data cross-referenced with platform reports...

Conversion paths (last 30 days, 473 conversions):

Path Conversions % Avg Value
Google (direct) 142 30% $89
Meta (direct) 118 25% $72
Google → Meta → Convert 67 14% $124
Meta → Google → Convert 48 10% $108
TikTok → Meta → Convert 31 7% $95
LinkedIn → Google → Convert 22 5% $186
TikTok → Google → Convert 18 4% $102
3+ platforms → Convert 27 6% $210

Key insight: Multi-platform customers are worth more. Single-platform converters average $80, while customers who interact with 2+ platforms average $131 — 64% higher.

The LinkedIn effect: LinkedIn rarely converts directly, but LinkedIn → Google paths have the highest average value ($186). LinkedIn builds awareness among decision-makers who later search on Google to buy.

Budget implication: Cutting LinkedIn would save $2,500/month but cost you ~$4,092/month in Google-attributed revenue from LinkedIn-initiated paths. Keep LinkedIn — it's the top-of-funnel driver for your highest-value conversions.


Cross-Platform Tips from the Pros

  1. Start every Monday with a cross-platform report. Normalize metrics across platforms (ROAS, CPA, CTR) to compare apples to apples. The agent does this automatically.
  2. Reallocate from worst to best ROAS channel. Check monthly. If Microsoft is 5.5x ROAS and LinkedIn is 1.7x, shift budget. Most marketers leave money on the table by not rebalancing.
  3. Don't judge platforms in isolation. LinkedIn might look expensive on last-click, but it initiates conversion paths completed on Google. Multi-touch attribution reveals the full picture.
  4. Coordinate launches. When promoting the same product across platforms, use consistent messaging, shared UTMs, and synchronized timing.
  5. Set platform-specific KPIs. Google = ROAS, LinkedIn = cost per SQL, TikTok = cost per video view, Meta = cost per purchase. Each platform excels at different things.
  6. Detect anomalies early. A broken pixel, a competitor bidding on your brand, or creative fatigue — catching these in 1 day vs 1 week saves hundreds to thousands.
  7. Multi-platform customers are your best customers. People who see your brand across multiple channels convert at higher rates and spend more. Cross-platform presence isn't waste — it's reinforcement.

FAQ

Is there a single MCP for all ad platforms?

Yes — this repo. It pre-configures the Synter MCP server with access to Google, Meta, LinkedIn, Microsoft, Reddit, TikTok, X, Amazon, Spotify, and The Trade Desk.

Can one AI agent manage multiple ad platforms?

Yes. The Synter MCP provides unified access to all major platforms. Ask one question, get data from all channels.

How does cross-platform budget reallocation work?

The agent compares ROAS, CPA, and other metrics across platforms and recommends moving budget from underperformers to top performers. You approve the changes before execution.

Does this replace individual platform agents?

This repo gives you access to everything. The individual platform repos (google-ads-agent, meta-ads-agent, etc.) are focused entry points with platform-specific documentation.

Can I launch campaigns on multiple platforms at once?

Yes. Describe your campaign once, and the agent creates platform-appropriate versions on Google, Meta, LinkedIn, or any combination you choose.


Related Repos

Every platform has a dedicated repo with deeper documentation:


License

MIT — see LICENSE for details.

Built by Synter · Get API Key · Documentation

About

Manage ads across all platforms with AI agents via MCP (Amp, Cursor, Claude Desktop, VS Code)

Topics

Resources

License

Contributing

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors