SEO Cartograph — Dashboard guide

Your first crawl

1. Download the desktop app — the Licenses page of the dashboard, Windows or Linux button. On first launch, Windows may show “unknown publisher”: More info → Run anyway.
2. Activate it with your key — it’s waiting for you on the same page (copy button). One key = one machine.
3. Run your crawl — enter your site’s URL in the app, click Crawl. Give it a few minutes depending on the size of the site.
4. The crawl flows up on its own to the dashboard — the Getting started page updates automatically, so leave it open.

With every new crawl of the same site, the history grows richer: score, pages, trend. Nothing is ever overwritten.

Choose your site, once and for all

At the top of the sidebar, the Site selector: you pick your site once, and every section — Performance, Editor, Netlinker, Crawls — follows. You never re-select anything when switching pages.

The Crawls section

Each site has its own card: audit score, number of crawls, pages, trend. When you open a site:

• the crawl history with the score curve over time;
• each crawl opens onto its detailed audit — issues found (titles, meta, internal linking, 404s, redirects…), the list of pages, the internal link graph;
• from a crawl, you can send issues to the Editor to handle them as tasks.

Performance, the cockpit

One project per site, seven tabs. Every curve is interactive and every time range is free: 7/28/90 days, last month, last quarter, or any dates you like.

Overview

The site’s KPIs over the period — clicks, impressions, average position, CTR on the Google side; sessions, conversions, engagement on the GA4 side — with comparison to the previous period. Including the traffic coming from AI engines (ChatGPT, Perplexity…) if your site gets any.

Clusters

Your pages grouped into semantic silos, pre-split by language/country if your site is multilingual. For each cluster: pages, clicks, impressions, position, CTR, conversions, coverage, internal linking.

• Deep dive: cluster curves, its pages, history rebuilt page by page, monthly snapshot.
• Persistent editing: rename, merge, move — your choices are locked, never overwritten by a crawl or an AI reorganization.
• Cluster SERP analysis: the real Google SERP on the cluster’s queries, with typed recommendations — quick win (page 2 → page 1), cannibalization, absent from the SERP, content gap, PAA questions — and your share of voice against competitors.

Keywords

Your rank tracker, on the real SERP (not just GSC data).

• Add keywords by hand, or from any area that displays queries: hover → star → bulk select → choose the group → send. A solid yellow star means already tracked.
• Each keyword has its target page: you see whether Google ranks the right page, with an alert if it doesn’t.
• Groups, position, 7-day change, best position, history as a curve.

SERP checks use up your monthly quota (see plans); Search Console positions, on the other hand, are unlimited.

Competition

Declare your competitors, then: the keywords × competitors grid (your position against each one, measured in the same SERP pass) and the ecosystem — who holds the top 10 across your keyword universe, day after day.

Explore

The microscope: 16 cross-views between crawl, GSC, and GA4 — pages gaining or losing, queries with no dedicated page, pages with no clicks, conversions by landing page, views by keyword. Each page opens in a detailed panel.

History

Every Monday, a snapshot is archived: positions, top 10, clicks. This is the project’s long-term memory and the source of the email alerts.

Settings

• The synced crawl: the latest by default, or pin a specific crawl — the whole cockpit aligns on it.
• Google connections: Search Console and GA4 (e-commerce included).
• Tracked competitors, subscription and billing, and for the Agency plan the white label.

The client report

Report button inside the project: a clean, printable page — KPIs for the period, clusters, tracked positions with a head-to-head against competitors, AI traffic.

White label (Agency plan): in Settings, set your name, your logo, and your color — the report comes out in your colors, with no Cartograph mention at all. Perfect for delivering to your own clients.

The Editor

The issues detected by the audit (or sent over from a crawl) become fixes to handle: a title to rewrite, a missing meta, a broken link… Each fix has a status — to do, done, ignored — and the whole thing exports. Proposed values can come from AI via MCP, but nothing is applied without your approval.

The Netlinker

Built on the Performance data: it knows your silos, your weak pages, your pages that deserve some link juice.

• Priority targets: the poorly linked pages that matter.
• Internal link suggestions (analysis or AI): anchor, source, target — you approve or reject. You’re the boss of the linking.

Licenses and plans

Search Console and GA4 data are never limited. The Licenses page: your key, its activation state, unlink/revoke a device, generate licenses for your team (Agency), and the app downloads. 14-day trial on the paid plans, cancelable in one click.

Billing

The Billing page: your subscription (status, trial, next due date), the history of your invoices with a downloadable PDF, and the Stripe portal to change your card or cancel. By email you receive: subscription confirmation, a receipt with each payment, a reminder 3 days before the trial ends, an alert if a payment fails.

Automatic alerts

Every week, if — and only if — something moves, you get an email “SEO Cartograph · your-site — …”: average position shifting, keywords entering or leaving the top 10, clicks breaking down, Google↔crawl discrepancies (404s that Google still visits, a stale crawl). Nothing to report = no email.

MCP: plugging AI into your data

MCP (Model Context Protocol) lets Claude — or any compatible assistant — query your Cartograph data directly: crawls, clusters, positions, SERP. Your assistant becomes an SEO consultant who knows your site by heart.

Connection

1. The Licenses page → MCP tab.
2. The server URL: https://app.tnedjar.com/mcp/cartograph
3. In Claude (settings → connectors), add this server: a connection screen opens, you sign in with your dashboard account and authorize. It’s OAuth — no key to copy, access scoped to your account, revocable.

What the AI can do

• Read: your sites and crawls, a crawl’s audit, a page’s details, your Search Console queries, your clusters and their KPIs, the linking map.
• Analyze: run a cluster’s SERP analysis, query the SERP on a keyword, inspect a URL the way Google sees it.
• Act — always under your control: propose cluster reorganizations (your locks are respected), put keywords into tracking, propose internal links or fixes — which land in the Netlinker and the Editor awaiting your approval.

Example prompts

• List my Cartograph sites and give me the score of each one’s latest crawl.
• Analyze the SERP for the “trail shoes” cluster and give me the 3 most profitable actions.
• Which pages have been losing clicks for 3 weeks? Cross-reference with the linking and suggest internal links.
• Put into tracking the 10 queries where I’m on page 2 with the most impressions, group “quick wins”.
• Inspect /blog/my-article and tell me why it isn’t ranking.

Quotas apply via MCP too (SERP checks, tracked keywords).

Quick FAQ

Where does my crawl data go?

The crawl runs on your machine. Only the results (pages, links, issues) flow up to your dashboard, nowhere else.

How fresh is the Google data?

Search Console: fresh to about 48h. GA4: the day before. SERP: real time at the moment of the check.

I’m switching machines.

Licenses page → unlink the device → activate the key on the new machine.

If I cancel, do I lose everything?

No. Crawls, history, and clusters are kept — if you come back, you find it all again.

A bug, an idea?

The bubble at the bottom right of the dashboard: it goes straight to Thomas.