Four tabs, zero answers
Analytics in one tool, revenue in Stripe, feedback in a third, errors in a fourth. The story lives in none of them.
Watching the demo workspace right now
All these charts. Still no answers.
Termind watches your traffic, product events and Stripe revenue together — then tells you what changed, why, and what to fix. Every claim comes with the evidence attached.
One script tag · EU-hosted · No card required
7,898events analysed
47people identified
15B2B accounts
€953MRR tracked
Live from the public demo workspace — the same pipeline, funnels and AI you get.
The problem
Analytics tools got very good at showing. None of them explain.
You are the analyst
Dashboards hand you charts and leave. Someone still has to notice the drop, segment it, and work out why.
AI you can't check
Chatbot summaries sound confident and cite nothing. You can't tell insight from hallucination.
What Termind does instead
Ask a question. Get the answer, the cause, and the receipts.
Answers with evidence, not vibes
Ask in plain English. SQL computes the metrics, Gemini explains them, and every claim carries the reference IDs it used — fabricated citations are stripped before you ever see them.
- Confidence and limitations shown
- Invalid references removed
- Model named on every answer
Why did activation fall this week?
Activation drop driven by mobile Google OAuth errors after release 1.14
HighOverall activation fell 9.8%. The decline is concentrated in mobile sessions using Google sign-in and coincides with a 640% rise in callback errors.
Confidence high (0.85)gemini-3.6-flash
m_activation_7ds_mobile_googlerel_1.14t_issue_0
Funnels and retention, computed not drawn
Signup to first value, drop-off by step, weekly cohort retention — every figure is a database query over your raw events, so it reconciles when you check.
- 2–8 step funnels with conversion windows
- Cohort grids with real counts
- Paths before and after any event
signed_up47 · 100%
onboarding_completed47 · 100%
project_created28 · 59.6%
19 people lost here · median 8 m to convert
Traffic joined to revenue
Connect Stripe in one click and every source, campaign and landing page is measured by the money it produced — not by bounce rate.
- First-touch and last-touch views
- MRR, expansion and churn per account
- Refunds and failed payments included
google
685 · 42.6%
producthunt
263 · 16.4%
twitter
232 · 14.4%
(direct)
159 · 9.9%
hackernews
139 · 8.7%
Visitors by first-touch source, joined to signups and MRR.
Installed before your coffee cools
One script tag gives you web analytics. One identify() call upgrades it to product analytics. The live debugger shows events arriving in seconds, so you never wonder if it's working.
- 7.8 kB, no npm package
- Localhost ignored by default
- Live event debugger built in
<script defer
src="https://www.termind.tech/t.js"
data-key="pk_your_key"></script>That is the entire install. 7.8 kB, no npm package, no build step.
Live event debugger
- $pageviewanonymous12:04:11
- signed_upidentified12:04:53
- onboarding_completedidentified12:06:20
- project_createdidentified12:09:41
- subscription_startedstripe12:31:07
How it works
Three steps. Ten minutes.
Paste one line
Page views, referrers, campaigns and landing pages start flowing immediately.
<script defer src="…/t.js" data-key="pk_…">Name your users
One call after sign-in connects behaviour to real people, accounts and their revenue.
termind.identify(user.id, { plan })Ask anything
SQL computes the numbers, the AI explains them — with the evidence attached.
“Why did activation fall this week?”Priced like a tool, not a platform.
Seven days free, no card. Every feature is on during the trial, because judging a product from a crippled version of it is not a fair trial.
Analytics
For a site where the question is where the traffic comes from.
€6.99/month€75.49 a year saves you two months
Start 7 days free- Unlimited websitesOne login, every site you run
- Visitors, sources, pagesWhere people come from and what they read
- Bounce rate and session timeMeasured from real activity, so an abandoned tab is not a two hour visit
- Live visitor countWho is on your site right now
- Channels, campaigns, entry and exit pagesIncluding where people give up
- Country, region and city
- Device, browser and system
- One script tagNo npm install, no build step
- CookielessNo consent banner needed for analytics
- AI crawler trackingWhich assistants read your pages, and whether it was to answer someone
- Data stored in the EU
- Signups, activation and funnels
- Revenue attribution
- Ask and alerts
Most people pick this
Complete
For a product where the question is which traffic turns into customers.
€11.99/month€129.49 a year saves you two months
Start 7 days free- Everything in Analytics
- Signups and activationThe full journey, not just the first click
- Conversion by sourceWhich channels bring people who actually sign up
- Revenue attributionStripe or Dodo, matched back to the source that produced it
- Search ConsoleWhich searches bring people who convert
- Assistant attributionVisitors sent by ChatGPT, Claude and Perplexity, and what they were worth
- AskPlain-English answers over every metric here, with a chart when it helps. Charts are checked against your data, never generated from a guess
- Alerts and weekly digest
- Your whole teamInvite people, set their access
Cancel any time from inside the app. Prices exclude VAT.
Why switch
The difference is who does the analysis.
| Traditional analytics | Termind | |
|---|---|---|
| Getting an insight | You interrogate dashboards until something looks off | The answer arrives written, with the cause isolated |
| Trusting the numbers | AI summaries with no sources | Every claim cites evidence refs; invalid ones are stripped |
| Revenue context | Analytics here, Stripe somewhere else | Behaviour and MRR joined per person and account |
| Setup | SDKs, event taxonomy workshops, a data engineer | One script tag, 7.8 kB, no build step |
| When data is thin | Charts render anyway and mislead | It says “not enough data yet” and waits |
Early words
Built with founders like you.
“I stopped opening four dashboards every morning. The briefing tells me what broke, where, and what it costs.”
“It caught a signup drop from one traffic source before I did — and linked the release that caused it.”
“Ask is the first AI feature I actually trust. It answers with my numbers and shows the receipts.”
Your next fix is already in your data. Go get it.
Paste one line, watch the first events arrive in the live debugger, and ask your first question today.
Start free