LET'S GET YOU GROWING

A few steps to
your first insight.

A practical guide to connecting your website and understanding what happens next.

STEP 01

Give your website a home.

Create an account, verify your email, and open My websites. Add a website name and its final homepage URL, such as https://example.com. Projects help organize multiple websites.

Use a public domain with HTTP or HTTPS and no custom port. Use the final URL: verification and checks do not follow redirects. The www and non-www domains are treated separately.
STEP 02

Show that the website is yours.

In My websites, copy your all-in-one installation block. Paste it once near the start of the server-rendered <head>. It contains a static ownership meta tag and a consent-aware tracker. Select Verify ownership after clearing your site/CDN cache.

For WordPress, download ToolMeadow Connect. Upload the ZIP through Plugins → Add New → Upload Plugin, activate it, and paste the same installation block in Settings → ToolMeadow Connect. Save, clear caches, then verify here. The plugin checks that the code matches your exact WordPress home domain.

Do not install both the plugin and a manual copy. Existing ownership stays saved; a public tracker ID or browser request cannot grant ownership. Each account uses its own proof.

STEP 03

One installation, clear visitor choice.

The default prompt offers Allow analytics and Decline analytics with equal controls. No analytics pageview is sent until approval. Analytics preferences lets visitors change their choice. A local-storage preference is written only after their action; if storage is blocked, the choice lasts only for this page. Withdrawal prevents later page loads; it cannot recall a pageview already sent. DNT and GPC override approval.

The script itself is downloaded before consent, so normal web-server request logs can exist. Configure your privacy notice accordingly. The built-in prompt is a simple opt-in control, not a compliance certification or universal consent-manager integration.

Already using a consent manager?

Use data-consent="required" instead of "prompt" in the manual snippet. Your consent manager must set window.toolmeadowConsent = true after genuine approval and dispatch the toolmeadow:consent event. Set the flag false on withdrawal and do not dispatch approval on later loads. Do not run a second competing banner. WordPress Connect uses the built-in prompt; custom CMP integration requires a separate manual installation.

One pageview per site ID per full page load is attempted, including when the current tracker is loaded twice. Remove old cached trackers: earlier versions cannot share the deduplication guard. No SPA navigation tracking. A successful browser fetch is not proof of ingestion; check the real Analytics report.

STEP 04

Keep a little peace of mind.

Open Monitoring for a verified website, enable the monitor, and choose an interval from 60 to 1,440 minutes. Checks run through the scheduler and queue; the first result may not appear immediately.

HTTP 200–399 is considered UP. Network failures and HTTP 401/403/429 are inconclusive (UNKNOWN); other HTTP failures require two consecutive checks before DOWN. Status-change emails depend on configured mail delivery. A stale or missing check is not the same as a confirmed website outage.

Current checks use one location, do not follow redirects, and do not inspect SSL-certificate expiry. There is no uptime SLA.
STEP 05

Find a useful next improvement.

Open Homepage SEO, choose your verified website and request a scan. Refresh the page to see the completed result.

The scanner checks 23 HTML and response checks: title, description, one H1, canonical presence, HTTPS, HTTP 200, robots noindex, image alt attributes and viewport. Complete-response scores summarize evaluated checks, not rankings. Truncated responses have partial findings and no complete score. For bounded multi-page crawling, issue categories, links and common-page comparisons, open Website audit. Its page cap is controlled by your plan; no JS rendering or full-web index is claimed.

Active-session analytics uses a random, consent-gated per-tab session token, rotated after 30 minutes. Only an HMAC is stored server-side with a page path and last-seen time. Heartbeats run every 30 seconds while visible; the live window is 90 seconds. Presence rows are cleaned every five minutes. This is an active-browser estimate, not unique people. CMP withdrawal should also dispatch toolmeadow:consent-denied.

A thoughtful approach to data.

The analytics table stores page paths, source domains, coarse device/browser categories and timestamps. It does not store visitor IPs, persistent visitor IDs, cookies or full user-agent strings.

That does not make the entire application free of personal data. Account information, authentication sessions, contact messages, queue payloads and web-server logs can contain personal information. Paths can also contain identifiers; exclude sensitive pages and keep identifiers out of tracked URLs.

Analytics raw/daily data currently uses a 30-day retention setting. Monitor checks retain 90 days. Cleanup requires the scheduler. Owners can export daily counts, delete analytics data, or delete a website and its related records.

Configure a suitable privacy notice and consent manager for your deployment. Read the privacy disclosure →

Know your room to grow.

Tool allowances reset on the first day of each UTC calendar month. Pageviews, monitoring checks and accepted scan requests each use their respective units. Accepted attempts count even if the target is unavailable.

A zero allowance means no usage. Changing a plan does not erase previously recorded usage. Paid subscriptions currently require manual activation, with no online payment or automatic billing.

Compare plans

AI Console: first-party discovery signals

Open AI Console from your workspace after verifying your website. Traffic uses your existing consent-aware tracker; no additional browser script or third-party data API is required. Reports use completed UTC days.

AI referrals match known source hostnames. Missing referrers stay unattributed. Crawler activity requires owner-uploaded JSON Lines logs, at most 500 rows and 256 KiB per file. Preserve unique request IDs; remove personal data before upload. Queries, IPs and full user agents are not stored by this module.

Reports include daily traffic, identifiable AI sources, imported crawler observations and downloadable CSV. Crawlers are classified by unverified user-agent claims. No search keywords, ranks, citation counts or model training claims are inferred. Raw-data retention and plan import quotas apply. This initial release does not include automatic log agents, goals, revenue attribution or scheduled email reports.

ToolMeadow assistantProduct help · automated

AI by Groq. Sending shares your message and up to 6 recent replies with Groq. No passwords or private data. Privacy
Want a representative? Choose WhatsApp below.

PRODUCT GUIDE

Hi! What are you trying to do with your website today?

Confirm this change