What's New

Version 1.2.0 is the release where ResetUI becomes an AI web designer for SaaS. You describe the page you want and it gets designed, with no template to browse and no style to pick: the brief is matched to a curated design automatically. Twelve page types are now first-class, so sign-in screens, pricing pages, contact pages, dashboards, articles and 404s each come out composed for their own purpose. The Free plan works for the first time, chat editing is open to everyone, and the design chat, the code export and the credit system have all been rebuilt underneath. A new $9 Starter plan now sits between Free and Pro.

• NewAdded twelve page types. Ask for a login screen, a pricing page, a contact page, an app dashboard, a waitlist, a portfolio, a blog index, an article, an about page, a careers page or a 404 and you get that page, composed for its purpose, drawn in the chosen design's palette, type scale, spacing and components.
• NewAdded a monthly Free allowance of 10 credits, refreshed automatically. That is one full design generation plus five chat edits every month, where a Free account previously had nothing and hit a paywall before seeing anything.
• NewAdded the Starter plan at $9 a month: 20 credits monthly, two active projects, full 20-section pages and code export in every format.
• NewAdded in-place upgrades from Starter to Pro, with the prorated charge shown before you confirm.
• NewAdded conversation memory to the design chat. It now remembers the whole project rather than the last few exchanges, with a fill indicator in the composer and automatic condensing of older messages that never interrupts the conversation.
• NewAdded whole-page awareness to the design chat. A single request can now edit several sections, reorder them, add one or remove one, because the chat sees the entire document instead of one section at a time.
• NewAdded version history for page-wide styling. Undo, redo and restore now roll back the page's shared styles alongside its markup, so restoring an older version returns the design exactly as it was.
• NewAdded page-shaped loading skeletons. The canvas now previews the silhouette of the page being built, so a sign-in screen shows a centred card and a data screen shows a sidebar and table while they generate.
• NewAdded four new comparison guides to the blog covering v0, Lovable, Bolt, Framer and the difference between an AI app builder and an AI design tool.
• NewAdded a cookie consent banner for visitors in the EU, UK and Switzerland. Analytics stays off until you accept, you can reject everything in one click or choose per category under Manage preferences, and a browser-level Global Privacy Control signal is honoured automatically. Your choice is remembered for six months.
• FixedFixed framework exports arriving incomplete. Long pages could be cut off mid-file and still download as though they had finished. Exports are now checked for completeness and validity before they are handed over, and a failed conversion refunds its credits instead of delivering a broken file.
• FixedFixed framework exports being desktop-only. React, Next.js, Vue and Svelte exports are now made responsive for mobile and tablet during conversion, with collapsing layouts, scaled headings, a working mobile menu and no sideways scrolling.
• FixedFixed chat edits that reported success while the canvas never changed. Edits are now applied against a stable identity for each section rather than its position on the page.
• FixedFixed replies in chat being cut off mid-sentence. An incomplete reply is now detected and retried rather than shown as though it were finished.
• FixedFixed the published product information that search engines and AI assistants read. Plans, credits, page types and export formats had drifted out of date, so the site was describing a version of the product that no longer existed.
• FixedFixed a Free account being able to delete a finished design and generate another one in its place, which made the one-design limit meaningless.
• FixedFixed a lapsed subscription sometimes leaving the account with no credits. It now returns to Free with its full 10 monthly credits.
• RemovedRemoved the one-time plan. There is now a Free plan that renews monthly, plus the Starter and Pro subscriptions. Existing one-time accounts keep full paid access.
• RemovedRemoved the chat paywall. Editing your design in chat is available on every plan, at 1 credit per message.
• RemovedRemoved template picking from the composer. The prompt is the whole input now, and the design is chosen from your brief after the page's purpose, content and palette are known, which is a better decision than anyone can make from a thumbnail.
• RemovedRemoved the redundant design planning stage, which spent a model call every generation producing a design system that the chosen design specification had already decided.
• RemovedRemoved the retired website-capture engine and every unused path left behind it, cutting a large amount of dead code, two unused runtime dependencies and several unused database columns.
• RemovedRemoved the elapsed timer and the memory meter from the chat, which reported on the machinery rather than on the design.
• ImprovedImproved design quality by rewriting the generation prompt in first-person design terms. It previously described a captured website that no longer existed anywhere in the product, which quietly worked against every generation.
• ImprovedImproved generation speed and cost with strict prompt caching. The design specification is now byte-identical across runs so it is cached rather than re-sent, and a project-specific line that was silently breaking that cache on every single generation has been removed.
• ImprovedImproved accessibility of every generated page with an explicit output contract covering heading order, real form labels, focus styles, alt text and colour contrast.
• ImprovedImproved the navigation on generated pages with a strict contract: no links to sections that do not exist, and a call-to-action that always contrasts with the bar it sits in.
• ImprovedImproved how replies appear in chat. The answer is held back until the page has finished updating, then typed out, so the design and the explanation of it no longer arrive out of order.
• ImprovedImproved the framework export prompt with per-stack rules for React, Next.js, Vue and Svelte, so converted files follow each framework's own conventions instead of a generic translation.
• ImprovedEvery plan now generates on the same fast model and produces the identical design from the identical prompt. When the shared slots are busy, the queue runs Pro first, then Starter, then Free.
• ImprovedMade the standard HTML export free and unlimited on Starter and Pro. It is deterministic assembly with no AI behind it, so there was nothing to charge for. Framework conversions to React, Next.js, Vue and Svelte run a real conversion and cost 2 credits. Export stays locked on Free.
• ImprovedSimplified credit pricing to one price list that is identical on every plan: 5 credits per generation, 1 per chat message, 0 for an HTML export, 2 for a framework conversion. Plans differ by allowance and permissions, never by price.
• ImprovedImproved the export dialog with clearer stack names in a more sensible order: HTML + CSS + JS, NextJS, React + Vite, Vue and Svelte.
• ImprovedImproved a failed or cancelled generation so it no longer occupies a Free account's single project slot. Only a finished design counts against the limit, so a failure never blocks the next attempt.

Version 1.1.0 closes a security gap that let a global flag unlock paid features for every visitor, replaces the free-text prompt field with a required design selection, and rebuilds the design generation pipeline around an upfront planning stage. It also ships a new blog and a full landing page content and design pass.

• NewAdded a blog covering AI web design topics, with shared structured data so every page reports consistent site metadata to search engines and AI crawlers.
• NewAdded a new premium design directions section and a deeper product walkthrough section to the landing page.
• FixedFixed a security gap where a global development flag could unlock export, premium templates, and unlimited projects for every visitor if left enabled in production.
• FixedFixed the sitemap missing the pricing and blog pages, and updated it to report each page's real last-updated date instead of the deploy time.
• RemovedRemoved the free-text prompt field from the project composer. Every generation now starts from a design chosen in the gallery, which sets both the look and the page structure.
• RemovedRemoved the Free-plan-only template requirement. A design selection is now required to generate on every plan, not only Free.
• RemovedRemoved five underperforming free templates, tightening the free template library to a stronger, more consistent set of designs.
• ImprovedImproved the landing page with a full content and design pass across the hero, pricing, FAQ, testimonials, and footer, plus a reordered page structure.
• ImprovedImproved the design generation pipeline with a new planning stage that decides the layout archetype, color palette, type scale, and section plan before generation runs, so the model executes a settled design instead of inventing structure while writing.
• ImprovedImproved AI crawler coverage in robots.ts by explicitly naming eighteen AI crawlers instead of relying on a single wildcard rule.

Version 1.0.9 introduces a one-time $9 Trial plan, adds a post-generation feedback modal and five new premium templates, retires the design view and the beta request access flow, and strengthens premium template discovery along with subscription renewal and plan downgrade handling.

• NewAdded a one-time $9 Trial plan with 15 credits and no renewal, unlocking custom prompts, chat edits, production code export, and the premium template library.
• NewAdded a feedback modal after generation so you can rate the result and tell us exactly what to improve.
• NewAdded five premium templates, bringing the premium library to ten curated designs.
• RemovedRemoved the design view from the project workspace, leaving a single focused canvas and chat workflow.
• RemovedRemoved the in-beta request access flow so sign-up now goes straight to registration.
• ImprovedImproved premium template discovery with a dedicated Premium filter and clearer premium tagging in the template gallery.
• ImprovedImproved subscription renewal and plan downgrade handling with hourly billing reconciliation, so a lapsed subscription is corrected even when a billing webhook never arrives.

Version 1.0.8 adds a product demo video to the landing page, moves gallery and case-study media to WebP, retires the basic template set, and delivers a substantial landing page speed and accessibility pass.

• NewAdded a product demo video to the landing page showing the complete flow through to the finished design.
• FixedConverted template thumbnails and case-study media from PNG to WebP for noticeably smaller downloads at the same visual quality.
• FixedResolved a dashboard bug affecting the workspace experience.
• RemovedRemoved the eight basic templates, leaving a tighter curated library of standard and premium designs.
• ImprovedImproved landing page speed by deferring the demo video, case-study previews, and directory badges until they are actually needed instead of loading them during first paint.
• ImprovedImproved landing page accessibility with WCAG AA contrast on muted label text, corrected heading order, and keyboard-safe marquee sections.

Version 1.0.7 strengthens capture validation, adds dedicated error handling and new showcase content, expands the premium template library, and introduces product metrics tracking.

• NewAdded a dedicated 404 page for unmatched application routes with a clear recovery path.
• NewPublished five case studies covering real-world product design workflows and outcomes.
• NewAdded five premium templates with production-ready layouts and reusable design patterns.
• FixedAdded capture URL validation to reject malformed or unsupported targets before pipeline execution.
• ImprovedImplemented metrics tracking for product usage, conversion events, and application performance.

Version 1.0.6 removes brand accent changes, improves the dashboard experience and generation retry flow, and fixes generation failure bugs.

• FixedResolved generation pipeline failures that could interrupt jobs or leave projects in an inconsistent state.
• RemovedRemoved the brand accent customization system and its related design-processing workflow.
• ImprovedRefined dashboard navigation, project-state feedback, and interaction flow for a more responsive workspace experience.
• ImprovedStrengthened post-failure retry handling with isolated attempts, safer state recovery, and reliable credit accounting.

Version 1.0.5 removes automatic account-related email messaging and strengthens signup fraud protection.

• NewAdded fraud and abuse detection for signups, with automatic protection against suspicious multi-account activity.
• RemovedRemoved automatic account-related emails so users are no longer sent lifecycle messages by default.

Version 1.0.3 introduces design templates and prompts, expands export to multiple front-end stacks, and makes the user's design direction a required part of every generation.

• NewDesign templates and reusable design prompts
• NewFramework export options for Vite React JSX, Next.js App Router TSX, Vue, and Svelte alongside the existing HTML, Tailwind CSS, and JavaScript export
• NewCombined URL + design prompt composer on the landing page
• FixedProject creation no longer starts a generation without design instructions
• RemovedLegacy CSS design-output support
• ImprovedPrompt-and-URL creation flow across onboarding and the dashboard
• ImprovedConsistent product copy that explains how source context and user direction work together

Version 1.0.2 makes generation and chat feel live with skeleton loading states and streaming code output, tightens concurrency and limit handling, improves design quality, and removes design themes completely.

• NewSkeleton loading placeholders for planned sections while generation streams in
• NewLive chat streaming that shows phase labels and generated code output in real time
• FixedConcurrency handling across generation, chat, and export flows
• FixedPlan, credit, and project limits now fail or refund more reliably
• RemovedDesign themes are completely removed
• ImprovedDesign generation quality
• ImprovedGeneration and editing speed

Public Version-1.0.1 adds key editing controls, ships version history and direct signup, fixes chat-to-design behavior, and improves the design generation experience.

• NewUndo support for quickly reverting recent design changes
• NewRedo support for restoring reverted design changes
• NewVersion history for reviewing and returning to previous design states
• NewDirect signup flow for faster access without the beta request step
• FixedChat to design flow now applies requested changes more reliably
• ImprovedDesign generation quality, consistency, and iteration speed

Beta Version-1.0.0 improves the core model output quality, strengthens security and SEO, refines the user experience, and raises the overall design quality of generated pages.

• NewSecurity across authentication, generation jobs, and exported output handling
• NewSEO for generated pages, including better structure, metadata, and semantic output
• FixedModel output quality issues for cleaner layouts, stronger visual hierarchy, and better spacing
• ImprovedUX issues across the generation, preview, and export flow
• ImprovedDesign quality issues with more polished spacing, typography, color balance, and responsive behavior

The generation pipeline now processes every section of the page independently and in parallel, reducing median generation time. AI chat lets you iterate on any generated design without re-running the full pipeline.

• NewAI chat interface for post-generation iteration (1 credit per message)
• NewProduction-ready Code export (2 credits per export)
• NewUsage dashboard showing credit balance, transaction history, and per-feature consumption
• FixedStuck generation jobs (>10 min) now auto-marked as FAILED by the watchdog
• FixedAuto-scroll capped at 150 iterations to prevent infinite loops on tall pages
• ImprovedGeneration failure now shows retry status in real-time ('Retrying 1/2...')
• ImprovedExport modal reworked with framework and styling options
• ImprovedOnboarding flow streamlined to 3 steps

Introduced side-by-side comparison for generated designs, plus multiple design direction presets so you can control the visual language of the output.

• NewComparison slider for toggling between two design versions
• NewDesign theme presets: Editorial, Technical, Minimal, Bold, Modern
• NewProject dashboard with card grid, status indicators, and quick actions
• NewSettings page with account management and subscription overview
• FixedOccasional blank sections when source page used heavy JavaScript rendering
• FixedCredit balance not updating in real-time after export
• ImprovedCode output formatting with consistent indentation, Prettier-compatible
• ImprovedSection rendering pipeline more resilient to malformed third-party HTML

Source-aware generation, full-page capture, section-by-section analysis, and initial code export. Selected users only.

• NewWebsite submission and full-page capture via headless browser
• NewAI-powered section analysis: brand extraction, copy preservation, layout inference
• NewSingle-file production-ready code export
• NewKinde authentication with Google and GitHub sign-in
• NewCredit-based usage system with plan enforcement
• NewDodo Payments billing integration for Starter and Pro plans

Internal-only build used for testing the AI pipeline and capture system. Not publicly accessible.

• NewInitial Inngest-based async job pipeline
• NewPuppeteer-based page capture with viewport scrolling
• NewOpenAI integration for design analysis
• FixedSection deduplication in long-scroll pages