Same capabilities, two doors, one credit ledger.
An MCP server so your agent can pull a teardown mid-conversation, and a REST API for everything scripted. Access is part of every plan rather than an upsell, because every serious tool in this category already bundles it.
live todayEvery tool below is callable now: mint a key from the cabinet and use it. The tool list and the credit costs are fixed and will not change on you. Plan access is listed on the pricing page.
Tools and cost
What each call does, and what it costs.
A credit is one ad analysed once, for your organisation. Reading and searching what you already analysed is always free, and so is re-tagging after we change the question bank.
| Tool | What it does | Credits |
|---|---|---|
| teardown_brand(domain, country, max_ads) | Collect and analyse a brand's live ads. Returns tags with confidences, traction and landing-page leaks. | 1 per new ad |
| compare_brands(domains[]) | The niche view: norms, differences and the gaps nobody is running. | 1 per new ad |
| search_ads(filters) | Query ads already analysed, by tag, brand, traction or date. | 0 |
| get_ad(id) | The full record: every label with its confidence and the evidence text behind it. | 0 |
| landing_match(ad, url) | Run the ad-to-page check on anything, including a draft of your own. | 1 |
| preflight(draft, url, niche) | Policy risk, landing-page match, and similarity to what already has traction in the niche. | 2 |
| audience(brand) | Set, reached and implied audience, plus a suggested test structure. | 5 |
| get_report(id) | A rendered report. | 0 |
Connecting it.
Scripts and agents authenticate with an API key: shown once at creation, stored hashed, and rate-limited per key. Inside the cabinet you already have access, signed in, with no separate key to manage.
{
"mcpServers": {
"targetdemon": {
"type": "http",
"url": "https://api.targetdemon.com/mcp",
"headers": {
"Authorization": "Bearer td_live_YOUR_KEY"
}
}
}
}curl https://api.targetdemon.com/v1/teardowns \
-H "Authorization: Bearer td_live_YOUR_KEY" \
-H "content-type: application/json" \
-d '{"domain":"brand.example","country":"US","max_ads":100}'What comes back
Every classified field carries its probability and its status, so an agent can branch on an unclear answer instead of treating it as fact. The same claims ladder that governs the written reports governs the API responses.
What it will not do
No endpoint returns a competitor's spend, CTR, CPA or ROAS, because no such field exists in any source we read. There is no estimate of it behind a different name.