MCP: Cross-Platform Ad Intelligence (Meta + Google + TikTok)

TL;DR: An MCP server that gives any AI agent five tools to research competitor ads across Meta Ad Library + Google Ads Transparency Center + TikTok in one call: search_advertiser_ads, get_active_creatives, compare_competitors, track_new_creatives, search_ads_by_keyword. It wraps the upstream ad-library-intel actor, reshapes its dataset into small, deterministic, agent-ready JSON, and is x402 (USDC on Base) + Skyfire ready for token-less agentic payments.

How it works

This is a thin, AI-agent-facing wrapper. Each tool call invokes the upstream seibs.co/ad-library-intel actor with the right mode/inputs, then returns a clean { ok, count, items, summary, error } envelope - no megabyte media blobs, no nested archive cruft. Upstream PPE events (ad_record, creative_analysis, advertiser_tracking) pass through to the same run, plus a flat $0.005 mcp_tool_call per invocation.

Modes

Mode	What it does
list_tools (default)	Emit the tool catalog (names, descriptions, JSON schemas) + the agentic-payment descriptor. Free - no charge.
call_tool	Invoke one tool. Requires tool + args.
batch	Invoke a list of {tool, args} calls (max 10) in one run.

Tools

Tool	Use for
search_advertiser_ads	A competitor's creatives across all three platforms - format, copy, CTA, run-dates, active-vs-stopped, hook/offer tags.
get_active_creatives	Only the currently-running creatives - what a competitor is spending on right now.
compare_competitors	A cross-advertiser comparison: per-advertiser active/stopped, format lean, top angles, shared angles.
track_new_creatives	An advertiser's current set newest-first (the snapshot to diff; schedule the upstream actor for an automatic delta).
search_ads_by_keyword	Discover ads by keyword/term - which advertisers run a given offer/angle.

Call it

from apify_client import ApifyClient

client = ApifyClient("<YOUR_APIFY_TOKEN>")
run = client.actor("seibs.co/mcp-ad-library-intel").call(run_input={
    "mode": "call_tool",
    "tool": "compare_competitors",
    "args": {"advertisers": ["Nike", "Adidas"], "platforms": ["meta", "google", "tiktok"], "country": "US"}
})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item)

Agentic payments (x402 + Skyfire)

The tool catalog advertises an agentic-payment descriptor so AI agents know they can pay per call with no pre-provisioned token. Enable rails via actor env vars: X402_ENABLED=1 + X402_PAY_TO_ADDRESS (USDC on Base), and/or SKYFIRE_ENABLED=1 + SKYFIRE_SELLER_ID. Settlement is handled by Apify's platform when MCP monetization is enabled; this server just advertises the capability.

Pricing

Flat $0.005 per mcp_tool_call + upstream ad-library-intel pass-through (ad_record $0.004 / creative_analysis $0.008 / advertiser_tracking $0.010). list_tools is free.

Related

Upstream actor: ad-library-intel. Part of the Seibs.co intelligence portfolio.