Two surfaces.
One engine.
A dashboard for product & design teams to track UX quality across releases. An MCP server for developers to get UX feedback inside the IDE before commit. Same engine, same categories, same fix-ready output.
The quick scan is the taster. The run is the product.
A quick scan opens one public page and reads what it sees, free and without an account. Autopilot signs in to your product, uses it the way a new user would, maps the flows and scores every screen it reaches — then does it again on the next release and tells you what changed.
| Quick scan | Autopilot run | |
|---|---|---|
| Where it looks | One page, from outside | Inside the product, signed in, every screen it reaches |
| Flows | None | Sign-up, onboarding, checkout and the rest, with steps and detours measured |
| Evidence | The page's visuals | Visuals, the screen's controls and your real user data (GA4, Clarity, Firebase) |
| Cadence | Once | Every release, with a diff of what got better and what regressed |
| Findings | A sample, the rest locked | Every finding with evidence, priority and fix-ready code |
The living report that tracks your product.
Not a PDF you download once. A continuous surface that tracks every release, surfaces regressions, and delivers fixes your team can act on today.
UX Score at a glance
One number that captures your product's experience quality. Tracks release over release — see what improved, what regressed, what stayed flat.
5 dimensions, 9 core + 2 neurodiversity categories. Weighted composite backed by 200+ peer-reviewed studies.
Prioritized findings
Every issue ranked by severity with concrete evidence — contrast ratios, pixel measurements, element counts. No vague recommendations.
Fix-ready code
Every finding ships with code you can paste into your IDE. Not pseudocode — real CSS, HTML or JS, and design-token aware.
/* contrast 2.8:1 → 4.6:1 */ .cta-button { color: var(--canvas-white); background: var(--brand-deep); min-height: 44px; padding: 12px 24px; }
Token-aware. Outputs your var(--token) instead of hardcoded values.
Neurodiversity Lens
15–20% of your users think differently. The Neurodiversity Lens measures what WCAG alone can't — ADHD friendliness, dyslexia readability, autism predictability, sensory sensitivity and color-vision safety.
No extra setup. Scores come from your existing scan data — the engine reads your UI the way neurodiverse users experience it.
9 tools inside your IDE.
The agent that wrote the UI can ask Corexi to review it, pull UX rules, or trigger a scan. No browser tab, no copy-paste.
{
"mcpServers": {
"corexi": {
"url": "https://corexi.ai/api/mcp",
"headers": {
"Authorization": "Bearer crxi_live_..."
}
}
}
}Connects to what you already use
7 analytics providers for behavioral data. 6 IDEs for developer delivery. Zero new tracking on your product.
What makes it continuous
Autopilot runs on your schedule and after releases: signs in, walks the product, scores every screen. A run cut short by a restart is scored afterwards, not lost.
Weekly digest: score changes, regressing pages, priority actions. In your inbox, or in Slack or Teams.
Context accumulates — tech stack, tokens, behavioral signals, prior findings. Gets sharper over time.
The whole method is published: how a run is scored and how it signs in to your product.
You hear about it before your users do.
Three rules, checked after every run. You set the threshold and where it lands; nothing fires twice for the same run.
A run scores the product under the number you set. The one rule that fires on a partial run too: it is a statement about what was actually measured.
PX 63, below your threshold of 75
The product scored materially lower than the previous run, with the new high findings named in the message.
PX fell from 69 to 58 · 4 new high findings
The accessibility pillar fell since the last run — the one that carries WCAG and the rules an automated engine can verify.
Accessibility fell from 74 to 61
A regression rule only fires when the two runs can be compared: a run that reached far fewer screens than the last one is a smaller sample, not a fall, and it says so instead of alarming you. The Monday digest goes to the same places, and can be switched off.
Are you ready for agents?
Your product is no longer used only by people. Support bots, buying agents and assistants walk it on someone's behalf — and where an agent stalls, a person stalls too.
A customer's assistant fills your forms, places the order, changes the setting. An unlabeled button, a state that says nothing after a click, a flow that ends short — the agent stops there, and so would a person.
Operators, copilots and just-in-time interfaces are moving the question from "does my screen look right" to "can an agent use my product safely". Products that cannot be used by agents will not be recommended by them.
Flows reached without help, steps against the fewest needed, actions that did what they said, hand-offs to a human. Read from the same Autopilot run that scores your product; shown beside PX, with the evidence.
Corexi's Autopilot is itself an agent, so every run already holds the evidence: which flows it reached the end of without help, how many steps that took against the fewest a person who knows the product needs, which actions changed nothing, and where it had to hand over to a human for a code or a captcha. The score is arithmetic over those records — no model is asked to guess — and it sits beside your PX score with the lines a customer can check: “6 of 8 flows reached their end without help”, “Critical flow ‘Pay’ stopped at /checkout”.
The layer gets smarter every phase.
Each phase ships a clear, standalone capability. Value from day one; each upgrade compounds it.
Competitive Benchmark
BuildingCompare your UX Score against your industry average. See where you lead, where you lag, what peers are fixing.
Requires 100+ sites in the index. Growing daily.
Predictions
PlannedIndustry-aware trend forecasting — seasonal spikes, emerging patterns, proactive recommendations before issues hit metrics.
Built on scan history + behavioral signal accumulation.
Test + Compare
PlannedGenerate fix variants and validate them with synthetic AI users. Measurable outcomes, not opinions.
Run variant A vs B on a new layout before you ship.
Behavioral Snippet
LiveOne script tag, instant behavioral signals. No GA4 setup. Cookieless, anonymous, privacy-first.
Scroll depth, rage clicks, bounce signals into your UX Score.
Autofix Agent
FutureLow-risk fixes ship as Corexi-opened pull requests with one-click approval. High-risk stays behind human review.
One-click fix execution via Cursor, Claude Code, or terminal.
See it yourself.
We onboard every product personally: a 15-minute call, then your first Autopilot run. 14-day free trial, no credit card.