Tracetify — Competitor Ads, Growth Traces & Search Console
VerifiedFeaturedRemoteAbout Tracetify — Competitor Ads, Growth Traces & Search Console
Tracetify MCP is a marketing and SEO research server for clients such as Claude Code, Cursor, and Codex. It provides competitor growth traces, ad research, search performance, and site audits by forwarding tool calls to the Tracetify HTTP API; some actions require an API key and credits.
Key features
- 01Trace competitor growth and sources
- 02Query Search Console performance
- 03Audit technical site issues
- 04Research keywords, SERPs, and backlinks
- 05View competitor ads on Meta and Google
- 06Check AI citations of a brand
Use cases
- 01Compile sourced competitor growth research
- 02Find optimization opportunities in Search Console
- 03Check site links, titles, and content issues
- 04Analyze keyword rankings, backlinks, and competitor ads
MCP support
The MCP server exposes Tracetify research, search performance, and site audit tools to clients. It can run locally or connect to the remote service over HTTP and authenticates with TRACETIFY_API_KEY.
Configuration
Remote servers need no local install. Pick your client, then copy the config or install in one click.
https://tracetify.com/api/mcpclaude mcp add --transport http tracetify https://tracetify.com/api/mcp{"mcpServers":{"tracetify":{"type":"http","url":"https://tracetify.com/api/mcp"}}}Open in Cursor{"servers":{"tracetify":{"type":"http","url":"https://tracetify.com/api/mcp"}}}Open in VS Code{"mcpServers":{"tracetify":{"type":"http","url":"https://tracetify.com/api/mcp"}}}{"mcpServers":{"tracetify":{"serverUrl":"https://tracetify.com/api/mcp"}}}Run commands in a terminal; paste JSON into the client’s MCP config and restart. Servers that need sign-in will prompt on first connect.
https://tracetify.com/api/mcpAuth requiredTools
18| Parameter | Type | Required | Description |
|---|---|---|---|
| query | string | Yes | Domain or fragment, e.g. "weshop" or "weshop.ai" |
Details
From the project READMEtracetify-mcp
MCP server for Tracetify — trace how any product actually grew, read your own Search Console, and audit a site, without leaving Claude Code or Cursor.
Most competitive tools tell you where a product stands today. Tracetify reconstructs how it got there: the first mention, the quiet weeks, the directory wave, the launch spike. Twelve sources per trace, every claim linked to the page it came from.
Full setup guide and example questions: tracetify.com/mcp
Free agent skills
Start with Competitor Research and GSC SEO Optimizer for sourced competitor briefs and prioritized Search Console work. These free, MIT-licensed workflows use public sources or your own exports and do not require a Tracetify account. Browse the skill source. Connect this MCP server when you want additional Tracetify tools and data; authentication and applicable credit charges still apply.
Setup
Create an API key in the dashboard, then pick one of the two transports.
Claude Code — one command:
claude mcp add tracetify -e TRACETIFY_API_KEY=ttfy_... -- npx -y tracetify-mcp
Or skip the install entirely — nothing to install, not even Node:
claude mcp add --transport http tracetify https://tracetify.com/api/mcp --header "Authorization: Bearer ttfy_..."
Same tools, same balance; the server runs on our side.
Claude Desktop / Cursor take the same JSON (Cursor reads
~/.cursor/mcp.json):
{
"mcpServers": {
"tracetify": {
"command": "npx",
"args": ["-y", "tracetify-mcp"],
"env": { "TRACETIFY_API_KEY": "ttfy_..." }
}
}
}
Codex uses TOML in ~/.codex/config.toml — note the snake_case table name:
[mcp_servers.tracetify]
command = "npx"
args = ["-y", "tracetify-mcp"]
env = { TRACETIFY_API_KEY = "ttfy_..." }
Try asking
Your agent picks the right tool on its own — these are real questions, not placeholders:
- How did photoai.com get its first users? Cite the sources.
- My page ranks #12 for "ai headshot generator" — what should I change to reach page one?
- We just deployed — audit example.com and fix what you find.
- Where can I get my new SaaS listed for real dofollow links?
- What is competitor.com running on Meta and Google right now? Pull the hooks and landing pages.
If your agent ever reaches for the wrong tool, say trace <domain> and it will
come straight here.
Tools
Tools that cost credits quote the price first and wait for you to say yes. Reading existing reports never costs anything.
| Tool | Cost | What it does |
|---|---|---|
search_reports | free | Find existing growth reports by domain |
read_report | free | Read a full report — timeline & verdict follow your account's unlocks |
start_trace | 10 credits | Rebuild a competitor's growth from 12 sources (~60–90s); a fresh cached report comes back free |
get_trace | free | Poll a running trace |
unlock_report | 10 credits | Permanently unlock a report's full timeline, evidence & SEO detail (idempotent — never charges twice) |
gsc_overview | free | Your own Search Console: clicks, impressions, period comparison |
gsc_queries | free | Your real ranking keywords with position and CTR — find what sits at #5–20 |
gsc_pages | free | Your pages by search performance, including high-impression low-CTR ones |
site_audit_start | 3 credits | Crawl a site for broken links, missing titles, redirect chains, thin content |
site_audit_get | free | Poll an audit and read the issue list grouped by severity |
research_competitors | 2 credits | Who fights a domain for the same keywords, flagging which ones already have a report |
research_domain_overview | 3 credits | Estimated organic traffic and top keywords for any domain |
research_keyword_volume | 2 credits | Google Ads monthly search volume, CPC, 12-month trend and keyword difficulty (0–100) for up to 10 keywords (per market or worldwide; difficulty is US-based when worldwide) |
research_serp_breakdown | 3 credits | Who holds the Google top 10 for one keyword in one market: per position the domain, homepage vs inner page, domain rating (0–1000) and estimated monthly organic traffic — the weakest slot stands out |
research_backlinks | 3 credits | Referring domains, authority and anchor texts |
research_competitor_ads | 3 credits | The ads a domain is running on Meta (Facebook/Instagram) and Google right now — copy, CTA, landing page, creative URL, platforms, start dates |
research_brand_lookup | 12 credits | How AI assistants cite a brand: platforms, mentions, associated entities |
backlink_directories | 8 credits | Hand-verified directories that actually give dofollow links (billed once per day) |
Everything draws from your Tracetify credit balance — the same balance the website uses. No seats, no per-tool add-ons. Top up at tracetify.com/pricing.
What this package does
It is a thin protocol adapter: MCP tool calls in, Tracetify HTTP API calls out.
No business logic, no data of its own — your API key never leaves your machine
except as an Authorization header to tracetify.com.
License
MIT
Comments
New to MCP?
From what the protocol is to installing your first server, these answers cover most first questions.
An open protocol that defines how AI applications connect to external tools and data in a uniform way. Clients talk to the model; MCP servers provide tools, resources and prompt templates. They exchange standard messages, so one server can be reused by many clients.