How it works

One script in <head>. Live data inside a minute.

TagDrishti runs independently of your tag manager. Your developer pastes one async tag, the first beacon lands within seconds, and you start reading real-user tag fires from the dashboard. No rebuild, no CI, no GTM changes. On purpose.

Start the 14-day trial or run the free audit first →

# 14 days free · no card · one developer ticket · you keep your data

# install_to_first_event

Three steps. Most of it is waiting for coffee.

No SDK install, no schema drafting, no kickoff call. One async script in your <head>, six detection layers behind it, and an alert pipe wired to whatever channel your team already lives in.

Paste one tag in <head>

Your developer drops a single async script tag into your site’s <head>. Not through GTM, on purpose. If GTM is the thing that broke (blocked by an ad-blocker, knocked over by a consent change, killed by a bad container publish), the monitor that watches it has to keep running. That separation is the entire point of having a monitoring layer.

<script async src="https://cdn.tagdrishti.com/monitor.js"></script>
<script>
  window.TAGDRISHTI_CONFIG = {
    tenantId: "td-your-id",
    apiKey:   "td_live_xxxx"
  };
</script>

Real sessions start streaming

Six detection layers watch every real visit: network beacons, data layer, Consent Mode v2 state, console errors, GTM container events, vendor SDK hooks. 80+ vendors auto-recognised. 135 columns per event flowing through Pub/Sub into Cloud Run, then into your workspace BigQuery dataset. The first session lands inside a minute, usually before the developer’s tab has finished loading.

Anomalies surface, alerts fire

When a tag stops firing, a consent signal flips, or a pixel starts returning HTTP 400, the rule engine routes the event to Slack, Microsoft Teams, PagerDuty, email, or a custom webhook, with session IDs, payloads, probable cause, and revenue at risk attached. White-label PDFs build themselves for client reviews. You hear about the break before your client does.

# why_not_a_quarterly_audit

Quarterly audits catch problems after the quarter.

Synthetic checks miss the breakages that only show up in real browsers: consent banners, A/B tests, SPA route changes, geo-specific failures. TagDrishti watches the production traffic you already have, so you find out at 02:14 instead of next Tuesday.

audit / synthetic

Spot checks · days per cycle.

A bot logs in, runs through the same flow it ran last quarter, files a deck. By the time you read it, the container has shipped twice and broken once.

Spot checks, days or weeks per cycle

Headless browser, not real users

Consent banners and A/B tests rarely fire

Findings stale on the day they ship

Blind between audits

You hear about the break from finance

tagdrishti

Every real session. Continuously.

Same browsers your customers use. Same consent state. Same SPA routes. The check is the session, not a synthetic robot pretending to be one.

Every real session, in real time

Runs in the browsers your users run

Sees consent state, experiments, SPA routes

Current state, queryable in BigQuery

Six detection layers on every fire

Alert fires with the session ID attached

install once, watch forever

Paste the script. Read your first session.

14-day free trial. No credit card. One async tag in <head>, data visible in the dashboard inside a minute. If your developer can paste, you’re done.

Start free trial Run free audit first

developer install in <head>no GTM changesfirst event in < 60sGDPR · DPDP · CCPA