Contrast maths, accessibility audits, critique and heuristic methods, research plans, copy review, handoff. Connect it to your AI assistant with one web address and it does these properly instead of guessing. Try the checks below first. New to this? Start here, or see how to connect Claude, Cursor, VS Code, Codex or ChatGPT.
WCAG 2.2 ratios for text and UI pairs, APCA lightness contrast, the nearest colour that passes, and a colour-blind check that finds palette colours which merge under protanopia, deuteranopia and tritanopia.
Free: 60 runs a day. API: POST /api/contrast. MCP: check_contrast.
Reads one page's HTML and names the WCAG 2.2 criterion behind each problem: missing alt, unlabelled fields, vague links, empty buttons, skipped headings, zoom locks, autoplay, tiny targets. Static markup only, and it says what it cannot see.
Free: 20 runs a day. API: POST /api/a11y-page. MCP: audit_page_accessibility.
Nielsen's ten heuristics as a worksheet for a screen or flow, with keyword and markup signals that point at where to look, review prompts for each heuristic, and the 0 to 4 severity scale. The judgement stays with you.
Free: 20 runs a day. API: POST /api/heuristics. MCP: heuristic_worksheet.
Reads a page's markup and CSS for the AI design defaults: purple gradients, Inter everywhere, pill buttons with arrows, eyebrow labels, emoji headings, glow blobs, identical card grids and stock copy. Signals, not a verdict.
Free: 10 runs a day. API: POST /api/design-slop. MCP: detect_design_slop.
Reviews buttons, errors, empty states, tooltips and toasts against what each one needs to do, checks the set for consistency, and runs the AI-writing tells over all of it.
Free: 30 runs a day, 20,000 chars each. API: POST /api/ui-copy. MCP: review_ui_copy.
Sorts critique notes into observation, opinion, solution, vague and question, says why the weak ones are weak, and gives a skeleton (what I see, what it does to the person, a direction) to rewrite them.
Free: 30 runs a day, 20,000 chars each. API: POST /api/feedback-lint. MCP: lint_design_feedback.
A critique agenda for the stage your design is in: framing statement, timed rounds, the questions that fit the stage, roles, and the feedback to turn away until the next stage.
Free: 30 runs a day. API: POST /api/critique-plan. MCP: plan_critique.
Lints interview and survey questions for leading wording, two questions in one, hypotheticals, closed questions in an open interview, loaded words, jargon and unlabelled scales, with a rewrite for each.
Free: 30 runs a day, 20,000 chars each. API: POST /api/research-lint. MCP: lint_research_questions.
SUS scoring with grade, percentile and interval; how many test participants find how many problems; A/B test sample size and duration; survey sample size. Each answer shows its formula.
Free: 60 runs a day. API: POST /api/usability-stats. MCP: usability_stats.
Lints a product analytics event list for naming drift, vague events, UI-element names, personal data in properties and near-duplicates, or drafts a starter plan for a SaaS, shop, marketplace, content or mobile product.
Free: 30 runs a day, 20,000 chars each. API: POST /api/tracking-plan. MCP: lint_tracking_plan.
Runs every foreground token against every background token from a CSS variable file or tokens JSON, with nearest passing colours and colour-blind merges. Free up to 30 colours; Pro takes 200.
Free: 10 runs a day, 8,000 chars each. API: POST /api/contrast-tokens. MCP: check_contrast_tokens.
Checks target sizes and spacing against WCAG 2.5.8 (24 by 24 with the spacing exception), 44 for iOS and the web, and 48 for Android.
Free: 60 runs a day. API: POST /api/touch-targets. MCP: check_touch_targets.
Lints design token names: mixed separators, hue names where roles belong, values in names, location names, duplicate values, abbreviation drift and irregular scales.
Free: 30 runs a day, 20,000 chars each. API: POST /api/token-names. MCP: lint_token_names.
Scores a DESIGN.md (the brand and design context file models read before building): missing sections, no hex values, text pairs that fail contrast, vague adjectives, no never-list, too long.
Free: 30 runs a day, 20,000 chars each. API: POST /api/design-md-lint. MCP: lint_design_md.
Every check, every method and every reference guide, right here or in your AI assistant, with a daily cap per tool (10 to 60 runs, listed above). Enough to check a palette, audit a page, or sort a pile of critique notes. Exactly what each plan includes.
/mcp.