00·Synthetic UX intelligence

Find your UX bugs before your users do.

AgentLens runs 100+ LLM personas through your site like real visitors — clicking, hesitating, bouncing — and tells you exactly where they got lost. No participant recruiting. No waiting for traffic.

  • Heatmap a page in ~90 seconds
  • Bring your own audience personas
  • v1 vs v2? Battle mode picks a winner
  • Opus-written exec report, not raw charts
example.com/checkout
● live
ambiguous CTA
bounce
converts
Agents
100
Steps
1,284
Drop-off
23%
Goal hit
58%
~90s
to your first heatmap
1,000
agents per run, max
8
persona dimensions
$0
to start — 100k free credits
01·How it works

URL in, heatmap out. Three steps.

Reproducible with a seed, streamed live as it runs, and you only pay for the steps that actually completed.

01
Drop a URL, set a goal

"Buy a razor under $30." "Sign up for the trial." Or no goal — let the agents wander like real visitors.

02
Pick your crowd

Use our diverse default matrix, or upload markdown personas of your real audience. We screen out the fake ones.

03
Read the report

Heatmaps, drop-off paths, persona-specific failures, and an exec summary your team will actually read.

02·Capabilities

What a UX researcher would notice — at 100× speed.

The same insights you'd get from a five-week usability study, on the same day you push the deploy.

Every hot spot, explained

Click, confusion, attraction, repulsion — four overlapping channels, each dot annotated with the agent's exact thought. No more guessing why users clicked.

Personas that behave like real users

Eight dimensions of realism — confidence, device, accessibility, intent. Mix our defaults, or bring your own audience.

An exec report your team will read

Top drop-offs, persona-specific failures, prioritized fixes — written in plain English, not just charted.

Upload your real audience

Drop in a markdown file describing your target users. We screen for nonsense and prompt injections before any agent reads them.

Watch it think, in real time

Every agent step streams to your dashboard as it happens. No five-minute black-box waits.

Drill to the moment of confusion

Click any heatmap dot or report screenshot. Zoom in. See exactly what the agent saw — and what they thought about it.

03·Personas

Synthetic users that behave like real ones.

A nurse on her second 12-hour shift browses differently from a regulatory buyer with a spec sheet. We model eight dimensions of that difference — and you can write your own.

Sleep-deprived NICU nurse
mobilein a hurryskims headlines

I need to see shipping date before I scroll. If it's not visible by the second tap, I'm gone.

First-time gift buyer
desktopresearch-heavyanxious

I'll open every product link in a new tab and compare before I commit. Reviews matter more than price.

Screen-reader user
a11ydeliberatefluent

Unlabelled buttons make me back out. If your alt text is empty, your conversion is empty.

Returning power user
expertfastloyal

I know where everything is. Don't redesign your nav without telling me — I'll fail upward.

Curious window-shopper
mobilelow intentbrowsing

I'm not buying today. But your interstitial newsletter popup just guaranteed I never come back.

Regulatory buyer
desktophigh stakesskeptical

I need PDF spec sheets and compliance docs before I add to cart. Marketing copy is noise.

04·The four channels

One screenshot. Four overlapping truths.

Most heatmap tools guess where users looked. We read what they thought. Every dot is sourced from an agent's reasoning trace — hover it to see the exact line that put it there.

  • Attention
    Where agents linger. Aggregated dwell + re-reads on the same element.
  • Confusion
    Hover loops, back-and-forth scrolls, the moment the LLM says "I'm not sure what this is".
  • Attraction
    Elements that successfully draw a click within the agent's first two steps.
  • Repulsion
    What makes them bounce — pop-ups, ambiguous CTAs, broken expectations.
mix-blend-mode: screen
06·Battle mode

Skip the six-week A/B test. Crown a winner by lunch.

Run two simulations of the same site with different changes. Same persona crowd, same goal. We diff the population, surface who flipped, and write the verdict line your stakeholders need — before a single real user sees the redesign.

  • Same Sara, two sites

    We pin the persona seed across both runs. The same sleep-deprived NICU nurse browses v1 and v2 — apples-to-apples, not a fresh random crowd each time.

  • A verdict, not a chart

    Headline first: “v2 beats v1 by +12pp on completion rate.” Then the persona buckets that flipped, the pages that improved, and the ones that got worse.

  • See exactly what changed

    Side-by-side heatmaps with a ⊕ diff overlay — lime where the redesign helped, rose where it didn't. Plus a line-level HTML diff for the receipts.

  • It's already on this account

    Every run you launch gets tagged. Drop two of them into a battle and we crunch the diff in seconds. No new setup, no new pricing.

00 · Battlerazor-checkout · v1 vs v2
v2 beats v1 by +12pp on completion rate.
v1· baseline
checkout-baseline

Current production checkout — control.

v2
checkout-new-cta

Replaced "Add to cart" with "Get yours today".

Metricv1v2Δ
Goal reached42%42%12pp
Avg confusion0.580.580.17
Dropoff @ /shipping38.0038.0026.00
Avg trust0.510.510.06
Personas where v2 won
mobile · in-a-hurry+24pp
desktop · researcher+11pp
a11y · screen-reader-8pp
07·MCPPro+

Run AgentLens from the chat you're already in.

Plug AgentLens into Claude Desktop, Cursor, or Claude Code with one command. Your AI assistant launches sims, watches progress stream live, and summarises the report — without you ever leaving the conversation.

Two commands to install
$ pipx install agentlens-mcp
$ agentlens-mcp configure
# paste your Pro key once. that's it.

Then drop one line into your editor's MCP config and your assistant gets eight new tools — launch, wait, summarise, compare, and friends.

  • No SDK. No client code.
    Your assistant calls the tools directly. You go from “what should we test?” to “here are the three things to fix” in one conversation.
  • Progress streams into the chat.
    Long-running sims surface live updates — 14/20 agents complete — through MCP's progress channel. No silent ten-minute waits.
  • Same API key. Same plan gating.
    Uses your Pro API key under the hood — every call respects your credit budget, your run history, and your seat limits. No new permission surface to learn.
Claude · agentlenslive
A
You

Run a 20-agent sim on staging.acme.com/checkout. Focus on mobile, and tell me what to fix before Thursday.

C
Claude

On it — launching the run.

toollaunch_simulation()
website_url:staging.acme.com/checkout
agent_count:20
objective:Complete checkout on mobile viewport
simulation_id = sim-7f3a… · 200-credit budget
toolwait_for_simulation()
6 / 20 agents complete
14 / 20 agents complete
19 / 20 agents complete
completed in 11m 22s · 184 credits used

14 of 20 mobile users dropped at the persona-selector modal. The three things blocking shipping Thursday:

  1. 01Modal traps focus — screen-reader users abandon at 81%.
  2. 02Primary CTA falls below the fold on iPhone 14 viewport.
  3. 03“Continue” is greyed out until the address validates, but the spinner never appears. Users think it's broken.

Want me to open a Linear issue with these and the screenshots?

mcp: agentlens8 tools active
07·Pricing

Four plans. One credit. Zero surprises.

Pay for what you ship, not what you might use. Annual saves 20%.

Hobby
For your first look
$0/ mo
Forever
  • 100 credits / mo
  • Balanced agent model
  • 1 seat
  • 7-day run history
  • Public share, watermarked
  • Custom personas
  • Battle mode
  • MCP server (drive from Claude / Cursor)
Plus
For solo practitioners
$47/ mo
$566 billed annually
  • 1,500 credits / mo
  • Flagship agent model
  • Custom markdown personas
  • Private share
  • 30-day run history
  • 1 seat — add more at $25/mo each
  • Extra credits at $0.05 each
  • Battle mode
  • MCP server (drive from Claude / Cursor)
Most popular
Pro
For UX & product teams
$159/ mo
$1,910 billed annually
  • 8,000 credits / mo
  • Flagship agent model
  • Custom personas + persona library
  • Battle mode — head-to-head comparisons
  • MCP server — drive AgentLens from Claude / Cursor
  • Priority queue (jobs run first)
  • API access
  • 1-year run history
  • 3 seats — add more at $25/mo each
  • Extra credits at $0.035 each
Most teams land here.
Enterprise
For organizations
Custom
from $1,499 / month
  • 40,000+ credits (negotiated)
  • Everything in Pro
  • Dedicated CSM + SLA
  • Audit logs + SSO / SAML
  • Priority support
  • Unlimited run history
  • 10+ seats included
  • Extra credits at $0.025 each
Annual contract only.
Need a one-off campaign? Pay as you go.
No subscription. No expiry. Stack on top of any plan.
Buy credits
500credits
$35$0.07 / credit
Most flexible
2,500credits
$129$0.05 / credit
10,000credits
$399$0.040 / credit
All plans include unlimited simulationsSwitch plans any timeSLA + uptime guarantee on Enterprise only
08·Get started in 90 seconds

Ship the report that already knows what your users will say.

100 free credits on Hobby — enough for a full demo. No card up front.