How our AI works
One engine.
Every signal, verified.
Corexi does not guess. It captures your live product, analyzes it with vision AI, verifies every accessibility claim with axe-core, fuses behavioral data from your analytics stack, and ships fix-ready code to your IDE. Here is how.
The pipeline
From URL to fix-ready code in under 60 seconds.
Capture
~3-8sPlaywright renders your product at desktop (1440px) and mobile (390px) viewports. Multi-state probes test focus indicators, modal dialogs, and form validation. Bot-protection bypass with fallback chains. No static screenshot, a living render.
Analyze
~8-15sClaude Sonnet 4 reads the rendered UI like a UX expert across 11 evaluation categories. axe-core runs in parallel, cross-checking every accessibility finding against WCAG 2.2 rules. 27 rule IDs verified. No hallucination passes unchecked.
Fuse
~1-3sBehavioral signals from up to 7 analytics providers (GA4, Clarity, Mixpanel, Amplitude, PostHog, Firebase, Hotjar) are merged with visual findings. The Pulse UX Engine computes weighted PX Score, neurodiversity scores, and confidence level. Each scan accumulates product context, compounding accuracy over time.
Ship
~2-5sEvery finding ships with fix-ready code that respects your design tokens. Choose your output: direct fix code, IDE-tailored markdown prompts (Cursor, Claude Code, VS Code, Windsurf, Replit), bulk export, or MCP delivery. Report with annotated screenshots and bounding boxes.
What makes it different
Six things no other tool combines.
Verified, not guessed
Every accessibility finding is cross-checked against axe-core. Verified claims carry a badge. LLM-only claims stay flagged. You always know the confidence level.
Gets smarter over time
Each scan feeds the engine six signals: tech stack, design tokens, growth goal, UI state, behavioral snapshots, and prior scan history. The layer learns your product.
Design-token-aware
Upload CSS variables, Tailwind theme, or Style Dictionary. Fix code outputs var(--your-token) instead of hardcoded values. Colors matched by perceptual distance (DeltaE <= 6).
Not a black box
Every score traces to a named research source: Nielsen Norman, Baymard, WCAG 2.2, Fitts, Miller, BDA, CUD. 200+ peer-reviewed sources, all auditable.
Proactive, not just reactive
get_ux_rules gives the agent UX patterns BEFORE writing code. review_code checks the result BEFORE commit. No other tool offers pre-coding UX guidance inside the IDE.
Neurodiversity as a first-class layer
Color Vision Deficiency and Dyslexia Readability are scored as dedicated scan categories. WCAG covers compliance. Corexi covers whether real people can actually use it.
By the numbers
Measured, not marketing.
200+
Research sources
NN/g, Baymard, WCAG 2.2, Fitts, Miller, BDA, CUD
11
Evaluation categories
9 core UX + CVD + Dyslexia
7
Analytics providers
GA4, PostHog, Clarity, Hotjar, Mixpanel, Amplitude, Firebase
9
MCP tools
Across 6 IDEs, live on cursor.directory
27
axe-core rules
Cross-checked on every accessibility finding
92.9%
Eval recall
Must-detect findings, CI-gated baseline
Honest about limitations
What we can and cannot do.
Dynamic content
Corexi captures initial render + multi-state probes (focus, modal, form validation). Infinite scroll and deep lazy-loaded content may need multiple scans.
AI accuracy
Strong on structural issues (contrast, spacing, hierarchy). Subjective design quality is harder. Every finding shows evidence so you can verify. Eval baseline: 92.9% recall on must-detect findings.
Authenticated pages
Currently public pages only. Authenticated scanning is on the roadmap.
Data handling
Screenshots processed in memory, encrypted at rest. Analytics via read-only OAuth. We never modify your analytics or codebase.
Try it yourself.
Paste a code snippet in the live demo or sign up to scan your product.