Example · find
Why did mobile checkout drop?
Rage clicks on Apply coupon rose after Tuesday's release. The button sits under the sticky footer below 390px.
Spot the problem, test the fix and ship the winner, without switching tools.
Cuttl puts heatmaps, session recordings and A/B tests in one tool.
See where visitors get stuck, test a fix, and see which version converts better.
How Cuttl works
Cuttl finds where visitors get stuck in recordings, turns the evidence into a test, and shows which version wins.
Spot
Test
Ship
/pricing
Click heatmap
Last 7 days
Issue · rage clicks
"Choose plan" doesn't respond on mobile
Draft experimentExperiment · from issue #128
Bigger "Choose plan" button
A · 50%B · 50%
A · Control
3.1%
B · Bigger button
4.2%
B has a 96% chance to beat control · guardrails green
Roll out BImpact report · issue #128
Conversion up 35% since rollout
Before3.1%
After rollout4.2%
Lift confirmed
You can Replay any visit
in any variant
Most teams pay for two tools.
Cuttl does both jobs.
Shows the problem
Tests the fix
One snippet · what you find becomes the test
Ask Claude or ChatGPT about your site.
Answers come with the recordings behind them.
Example · next test
What should we test next?
Moving Apply coupon above the footer. It hits 412 sessions a week, more than any other checkout issue.
Example · results
Is the new checkout winning?
Variant B leads with a 96% chance to beat control. Guardrails are green, and 4 more days reach the planned sample.
Connected through MCPAsk
Spot.
Heatmaps and recordings show where visitors get stuck.
Test.
Turn what you saw into an A/B test in one click.
Ship.
Roll out the winner and keep an eye on how it converts.
Everything you need, start to finish
From the first rage click to the winning variant live on your site.
Heatmaps
Session recordings
Rage clicks
Funnels
Issue detection
Test ideasA/B tests
Multivariate tests
GuardrailsImpact reports
Before and after
AI answersWhat you find feeds what you test
Anything you spot can become a test, and each test keeps the recordings that started it.
Test ideas arrive
with the evidence
Each issue becomes a test idea with its recordings, its heatmap and a count of the visitors it affects. One click drafts the test.
Hypothesis
Because we saw rage clicks on Apply coupon in 412 sessions, moving it above the footer will lift checkout.
R1R2R3H · page insight
Affects 18% of checkoutsDraft experiment
Draft created · Variant 1 ready
Every variant has
its own heatmap
A
B
See whether each
change worked
Impact report
Before3.1%
After3.1%
Lift confirmed
Works with the tools you use
Send findings and results to Slack, your AI assistant or your own code.
Claude, ChatGPT, Gemini
Ask about your visitors and tests in plain English, through Cuttl's built-in MCP server.
MCPREST API
Pull issues, experiments and results into your own dashboards and scripts.
/api/v1Webhooks and Slack
Get a message when an issue appears, a test finishes or a guardrail trips.
JSON · Slack presetExports
Download your raw data whenever you like. It belongs to you.
ExportsLocation without IPs
Cuttl reads region and city from Cloudflare, Vercel or CloudFront, so it never needs a visitor's IP.
Geo headersSingle sign-on
Team accounts with roles. Sign in through your company's identity provider over OpenID Connect.
OIDCOne plan, every feature
Plan includes
- Heatmaps, recordings and funnels
- A/B, split URL and multivariate tests
- Test ideas built from real visitor sessions
- An impact report for every test and release
- MCP server for Claude, ChatGPT and Gemini
- Per-variant heatmaps and recordings
Made for people
who fix things
If a page is yours to improve, Cuttl shows you what to change and whether it worked.
Launch tests from evidence, then watch per-variant replays to see why the winner won.
Experimentation teams
Hypotheses, A/B tests
"Which part of the funnel should we fix first?" Answered with recordings and the number of visitors each step loses.
Growth teams
Funnels, impact reports
See where your first users get stuck, fix it, and know whether the fix worked.
Founders
Heatmaps, change impact
Show clients the rage clicks on their checkout, then show them the conversion lift from the fix.
Agencies
Reports, A/B tests
Watch real people use the new layout in each variant before you defend it in a review.
Designers
Recordings, variant heatmaps
Check whether the feature you shipped moved the number, even without enough traffic for a test.
Product managers
Change impact, segments
Install Cuttl with one line of code
Paste it into your site's <head>. Heatmaps, recordings and A/B tests all run from this one snippet, so tests need no extra code.
<script async src="https://app.cuttl.io/t.js" data-site="SITE_ID"></script>
Questions people ask
Cuttl is a website optimisation tool that combines heatmaps, session recordings and A/B testing. It shows where visitors get stuck, helps you test a fix, and shows whether the change improved your conversion rate. All of it runs from one snippet.
When Cuttl spots a problem in a heatmap or recording, such as rage clicks on a button, it turns it into a test idea. The idea carries the recordings, the heatmap and how many sessions the problem affects. One click drafts the A/B test, and each variant gets its own heatmaps and recordings, so you can see why the winner won.
With two tools, you find a problem in one, rebuild the context in the other, then match the results up by hand. Cuttl keeps the evidence and the test in the same place, runs on one snippet, and comes with one bill.
An impact report shows what a change did to your conversions. When a test finishes or a change ships, Cuttl compares the conversion rate before and after, and tells you plainly whether the change helped, made no clear difference or made things worse.
Log the change instead. Cuttl's before-and-after check compares the page before and after the change against the rest of your site. If there isn't enough data to say whether it worked, Cuttl tells you so.
Bayesian by default, and you can switch to sequential or fixed-horizon tests. You choose the main metric and the smallest effect worth detecting before launch, and guardrails pause any variant that hurts your other metrics.
No. You paste one line of code into your site's <head>, or add it through your tag manager. Heatmaps, recordings and A/B tests all run from that snippet.
Cuttl never stores IP addresses, respects your consent settings before it records anything, and filters bots and AI agents out of every report.
Only if you connect them. Cuttl has a built-in MCP server, so Claude, ChatGPT or Gemini can answer questions about your visitors and your tests once you give them access.
Cuttl is a paid, fully hosted service: one plan with every feature and unlimited sites.