Developer Platform · MCP
One Model Context Protocol endpoint. Every fraud-defense and MMP capability — web scanning, in-app SDK ops, affiliate management, postback control, reports, IP lists, audience exclusion. Curated, scope-gated, and ready for Cursor, Claude Desktop, n8n, or your own backend.
Agent self-registration
No admin provisioning needed. Self-register, create accounts, add traffic sources, manage fraud reports, and operate campaign protection — all through MCP.
What you can automate
Curated tools shaped around real intents — add_affiliate, create_offer, upload_ip_list, run_report — not a 1:1 dump of 500 routes. Designed for LLMs to plan over.
web.*inapp.*mmp.*reports.*audience.*iplists.*trap.*site.*Website operations
https://clicktrust.cc/mcp/web exposes only the tools needed to operate web tracking accounts. MMP, in-app, audience and report tools are not reachable on it, whatever scopes the key carries — so a site-management agent can be given a key without widening its blast radius. Edits go live on the tag within 5 minutes. Accounts can be deactivated but not deleted. Writes are admin-only: every tool that changes a site requires the admin:web scope, which exists only on keys minted from Admin → API Keys. Partner-minted keys are read-only here.
Edit clicktrust.cc itself: with admin:site the agent can create and publish pages at any URL (HTML or markdown, inside the site nav/footer or full-page), replace existing marketing pages, write blog posts and upload images — site.pages.*, site.blog.*, site.assets.*. Published changes are live within a minute.
Read-only partner key: web:read, iplists:read, trap:read. Admin write key (Admin → API Keys): admin:web; admin tools take ownerEmail to name the customer. Catalog: /api/mcp/spec?surface=web. Config examples are under the “Website operations” tab below.
Quickstart
Sign up, generate a scoped MCP key from Developer → API Keys, paste it into your client config, restart the agent. Done.
Add ClickTrust to ~/.cursor/mcp.json so Composer + Agent can call tools.
{
"mcpServers": {
"clicktrust": {
"command": "npx",
"args": ["-y", "@clicktrust/mcp"],
"env": {
"CLICKTRUST_API_KEY": "ctk_mcp_xxxxxxxxxxxxxxxx"
}
}
}
}@clicktrust/mcp launcher bridges stdio (what desktop / IDE clients speak) to our hosted Streamable HTTP endpoint. It uses mcp-remote under the hood, so behaviour matches the upstream MCP spec exactly.Authentication
Pick the path that matches your runtime. Both feed the same scope-gated dispatcher; both can be revoked from your Developer → API Keys.
Best for backend agents, CI pipelines, n8n flows, and anything else that runs server-side. Mint a ctk_mcp_* key with the exact set of scopes the agent needs.
curl https://clicktrust.cc/mcp \
-H "Authorization: Bearer ctk_mcp_xxxxxxxxxxxxxxxx" \
-H "Content-Type: application/json" \
-d '{"jsonrpc":"2.0","id":1,"method":"tools/list"}'Spec-compliant authorization-code flow with PKCE. Best for Claude Desktop, Cursor, Continue, and anything else where you'd rather not paste a long-lived key.
{
"authorization_endpoint": "https://clicktrust.cc/api/oauth/authorize",
"token_endpoint": "https://clicktrust.cc/api/oauth/token",
"registration_endpoint": "https://clicktrust.cc/api/oauth/register",
"code_challenge_methods_supported": ["S256"]
}Scope vocabulary
Every tool declares the scope it needs. Grant only what the agent actually does.
web:readweb:writeinapp:readinapp:writeinapp:sdkmmp:readmmp:writemmp:postbackreports:readreports:writeaudience:readaudience:writeiplists:readiplists:writeTool reference
Generated client-side from GET /api/mcp/spec so docs are always in sync with the running server. Endpoint for this surface: https://clicktrust.cc/mcp
Example prompts
Drop these into Cursor, Claude Desktop, or any MCP-aware agent. Each one resolves to one or two tool calls.
"Create a tracking account for https://shop.example.com/landing, turn on lead-gen scanning and give me the snippet."
"Publish a new page at /pricing-india comparing our three plans in INR, in the site layout, with a meta description."
"Write a 900-word blog post on GAM ads.txt fraud and publish it with the tags gam and supply-chain."
"Switch account a1b2c3 to a multi-site tag and enable captcha blocking for high-risk clicks."
"Add an affiliate Acme Media to my SaaS Q4 offer with 20% rev-share and invite ops@acmemedia.com."
"Upload these IPs to my whitelist: 203.0.113.10, 203.0.113.11, 203.0.113.42."
"Run the weekly fraud report for account a1b2c3 and email it to alex@growthco.io."
"Show me the device blocklist for the iOS app and unblock the device that ends in …a4f2."
"Pull last 7 days of MMP clicks for affiliate Acme Media and tell me the attributed install rate."
"How many users would land in my Google audience-exclusion list for account a1b2c3 over the last 30 days?"
Free to try. Sign up, generate a scoped key from your Profile, paste it into your client config, and your agent has every fraud-defence and MMP capability one tool call away.