Changelog

Content is saved in one shape per channel, the way it will be posted

In the marketing phase, posts, card news, webtoons, and video scene tables the coach makes are saved under content/ in your project folder in a fixed shape. Channel previews that show each post as it will look, and pin fixes, read that shape.

• Every format leads with the answer. For a blog post, the search title, description, and first summary line say the same answer. For card news and webtoons it is the cover, and for video the first scene
• If search is how you bring in customers, the coach plans posts around search questions and first settles with you when numbers can be shown, who writes and who checks, and where each post sends the reader
• Pictures are drawn with the AI you use. If it cannot draw, the coach shows how to draw with your ChatGPT subscription. Without one, it makes text cards first

Ask AI about PurplePrint in one click

In Ask your AI at the bottom of the home page and in the demo, pick ChatGPT, Claude, or Perplexity and a new tab opens with the question already filled in. No more copying and pasting.

• Using another AI? You can still copy the question
• Each feature on the home page now has its own row: watch the coach recommend a first pick, and see how connecting your AI tool and starting works. Each row highlights the exact spot where that feature does its work
• The builder growth section now shows the real growth report. Click a feature such as builder level, leading stage, or the 7 signals and the screen points to it. The Hyper-Review section lets you walk through the report the same way, from the identity verdict to the handoff
• The user guide now has a section on what the coach does for you, feature by feature, including recursive checks, the order of development design, and task priority
• Adding several pins on the design map no longer covers the map. A long pin list scrolls on its own and keeps the newest pin in view. Product windows on the home page now show only the product. Walkthrough text and Back/Next sit outside the window, and only the highlight stays inside

The design map reads well in dark theme

Fixed table text in the data-relationship picture that blended into the background in dark theme. In the value chain view, the Now · After · Value chain switch is now a small tab row like the user journey, so it no longer looks like the step cards below.

Pricing shows the free trial and beta subscription side by side

The free trial (14 days · no card · 1 Hyper-Review) and the beta subscription ($13/month · 10 Hyper-Reviews a month) now each show what they include. The trial's single Hyper-Review no longer reads as a limit on the whole design coach.

• The Hyper-Review page shows a real review demo first, and asks you to sign in only when you choose to check your own doc

Choose how to fix your design right from the review report

If you reviewed an existing plan with Hyper-Review, pick a path under 'Improve with this report' in the report pane. Connect PurplePrint to the AI tool you use (Claude or ChatGPT desktop app, Cursor, VS Code) and paste the message, and the coach organizes your documents and fixes each finding.

• You can also download the report as .md and take it to your own AI. Turn off that service's setting that shares your data for training
• If you subscribe, you can still continue in the web design coach

A new project now starts by asking what you want to build

When you open a new project in the web coach, it now asks what you want to build instead of asking for a project name first. One or two lines are enough, and if you already have a business plan or planning doc, you can paste its text as is. The coach suggests a name from your answer.

• Dropping a file onto the chat no longer replaces the page. The coach tells you what works today: paste the text, or check the file in Hyper-Review
• When you press Start with my idea in the demo, the sign-in screen now speaks to first-time visitors

Connect once and you stay signed in while you keep using it

We fixed the issue where your AI tool (Codex, Claude Code, Cursor and others) asked you to sign in to PurplePrint again after about a month. The connection now stays alive as long as you keep using it, and one sign-in lasts at least 90 days. Only connections unused for more than 120 days close, for safety.

• AI tools you connected long ago can sign in again as they are, without removing and re-adding the connection. Your designs and account are kept

While you build, the coach checks only the part of the screen that changed

When you build in your AI tool, the coach no longer reopens and scans every screen after each small change. It checks only the part that changed and moves on, so you wait less and use less. The full pass over every screen, width and device happens once, right before you deploy.

• Tests that check your features still run as before

We tell you what to ask when you're unsure, and you can continue on another computer

If you're new, the coach mentions after its first reply and the early steps that you can ask "How do I use PurplePrint?" anytime. In your AI tool, once your first document is ready, the coach offers once to keep it in a GitHub repository. If you'd rather not, you just keep going.

• Once the repository exists, the coach gives you a one-line way to continue on another computer
• If you try to continue while PurplePrint is disconnected, it walks you through reconnecting (for newly connected computers)

While a Hyper-Review runs, the coach checks your situation once

When the coach runs a Hyper-Review for you, it prefills how you are building, your experience, what is hard right now, and your money and time from your materials. Just confirm, and the report's next actions fit your situation.

• When a review finishes in your AI tool, the coach shows how to fix your design documents with that report. If you have no documents yet, it first turns your planning materials into design documents

The document you are reading no longer reopens by itself

Fixed the open document reloading and jumping back to the top whenever the coach saved a different document. Now it only updates when that document itself changes, and keeps your place.

• On phones, finishing a step opens that step's document right away, with a button below to go back to the chat
• In the demo, a "Press to continue" bubble appears next to Send when it is your turn

We set out when the Company looks at your content, and you can choose in Settings

As a rule the Company does not open your web coach conversations, design documents or Hyper-Review inputs. Designated staff look only when you ask us to resolve something, when an outage or security issue needs it, or for quality checks, and every look is logged. This is in Section 6I of our Privacy Policy, effective October 24.

• You can opt out of viewing for quality checks under "Your data" in Settings
• You can ask for the viewing log through our contact

Follow a real design conversation from the start in the demo

Open the demo and, instead of the finished result, you see a founder's real conversation with the coach unfold one turn at a time. Press send in the founder's seat to see the coach's reply, and documents show what they looked like at that moment. When you reach the end, you can move on to the finished tour.

• When you finish a stage, a completion mark appears in the chat along with the documents made in that stage, and the finished stage glows once in the progress panel
• When you finish the whole design, a completion mark appears in the chat
• In the demo, a short card explains each scene the first time it appears (a choice, a saved document, a search, a new stage, a screen draft)
• Click a stage name in the progress panel to watch it from the start, and a finish card appears when you reach the end

Already have material? Start with a review

When you open a new project, "Start with a review" appears under the coach's first reply. Paste or upload your business plan or planning document, check it with Hyper-Review first, then bring the results back to the design coach to fix things. The line goes away once you start answering the coach.

See at a glance how your idea is protected

The same four promises now sit in "Before you start" on the home page, in the setup guide, and at the top of the privacy policy. AI processing is not used to train models, and external AI models are called only through zero data retention routes. With an AI tool, your conversations and documents stay on your computer, your designs are yours, and deleting means deleted.

• When you pick a project that is still at an early stage for Hyper-Review, we tell you first what a review at this point will show
• If Korean is anywhere in your device's language list, the web coach starts in Korean

Bring what you've already organized or built, and the coach starts from there

Paste an idea you organized elsewhere or a summary of the app you're building, and the coach first fills in the earlier documents from it, shows you, and saves after one confirmation. Pick where you're stuck and it goes straight to that step. If you want your existing design reviewed, it starts with a diagnosis.

• What you add while answering is merged into the earlier draft. You won't be asked to choose the same thing again
• If you say you don't know or don't mind, it's marked as undecided and the coach moves on
• The coach knows how many Hyper-Reviews you can run in 30 days and how many are left
• The account menu and the usage indicator under the input box now also show your remaining Hyper-Reviews and trial days

Current coaching version: v2.83.0.

Connecting the AI tool you already use is now one card in the chat

When an AI tool like Claude Code, Codex, or Cursor comes up, the web coach shows a single connection card in the chat. Copy the install prompt, restart the tool, and send the message, and you continue this project in that tool.

• If nothing is saved yet, tap 'Save what we've decided so far as documents' first. Only documents move with the project
• The pricing section on the home page now has a 'Start free for 14 days' button, so you can try the web coach right away without a card
• Your subscription status now stays correct even when payment notifications arrive late or out of order
• Our contact is now contact@purpleprint-ai.com. The Privacy Policy and Terms now name the operating company (Bringthehome Inc.) and our privacy officer (effective October 23)

Pick up right where you are in the AI tool you already use

If you already use an AI tool like Claude Code or Codex, the web coach now points you to connecting PurplePrint there first. In that tool the coach reads your project folder's code and documents directly and carries you from design through operations.

• Before you move, what you've decided so far is saved as documents first, so it carries over to the other tool
• If an earlier phase's document is missing, the coach drafts it from what you've already discussed, checks it with you, then moves on
• The end-of-trial card shows how many documents are actually saved, and if you're stuck you can book a free onboarding session right there
• The usage guide opens on the per-tool connection text. The copy button sits on top, and it reads better on phones

Signing out now signs out your other open tabs too

If the web coach is still open on a shared computer or in another tab of the same browser, signing out means that tab can no longer send messages to the coach. Sign in again to pick up where you left off.

• Screens on other devices where you are still signed in reconnect for a moment and carry on
• Hyper-Review intake limits now hold exactly, even when several requests arrive at once
• Pressing 'Previous version' while the coach is editing the same document no longer loses any version. If they overlap, you see the change first
• Your Hyper-Review reports open only for the account that made them. A copied address won't open for anyone else, so use 'Create share link' to show a report
• Even if a hidden instruction on a web page fools the coach, it won't open an address or fill another site's field with your conversation, documents, or contact details. When that's stopped, the activity line shows the site's name

Try the demo tour hands-on

The tour no longer just explains. It makes the exact button or sentence glow, and where you should try something yourself, like opening a document, tapping a document on the map, or adding a pin, it moves on once you do it.

• Conversation steps start with what the founder chose, then point to the coach's first choice and why the others wait
• You open the first three documents yourself to see how each one builds on the last
• In Hyper-Review you tap the component with the most issues, an issue bundled with others, the severity filter, and the message for each recipient. In My growth you open an earlier measurement
• The review screen now shows 'Fix my design with this review', so the tour also shows how a report flows back into your design
• On phones the tour card is smaller and moves out of the way of what it points to

Fix the gaps a Hyper-Review finds, right in your design

Once a review finds gaps, the next step is fixing your design documents. Press 'Fix my design with this review' above the report and the coach starts the fix in that project right away.

• The coach checks the report against your documents, maps each issue to the document and section to fix, and says what to fix first
• Documents change one at a time, and the coach shows each change and asks before applying it
• A review run without a project can be attached to the project you pick
• Reviews you receive in your AI tool also come with a one-time offer to fix the design

Updating keeps you where you were

When a new-version notice appears in an open web coach, you can now see what changed before updating. After updating, the documents, map, and demo tour step you were on stay as they were.

• Tap What's new in the notice to see what changed since you opened the page
• After updating, your open document tabs, the progress map, and the picture you chose open as before
• The demo tour has four new steps on the design map, and you can try pins in the demo too

The progress map no longer gets covered

When you reopen a finished design, the document map and the work to build should come first. We removed explanatory lines that repeated what the screen already shows, and the document card that covered the map now collapses.

• The document card starts collapsed to one line and opens when you tap a document on the map
• Drag its handle down or press the collapse button to fold it again
• Chip and icon explanations appear when you hover

Go straight from a Hyper-Review report to the design coach

After reading a report from an email link or My reviews, you often want to take what needs fixing straight into design. One button above the report now starts the design coach with that report already read.

• A review without a project gets a new project, with the report added as a document the coach reads first
• A review you ran from a design coach project opens that project right away
• For a review of a project made in an AI tool, you're told to continue in that tool

Deep research at each step now runs on the PurplePrint engines

When designing in an AI tool, the coach sometimes did deep research such as competitor analysis briefly, without the engine. It now starts the PurplePrint engine at the right moment, with the full research goals and evidence rules.

• Competitor, user, user journey, marketing and go-to-market research start on the engine at their step
• If you have Codex, it runs separately and only the result document comes back, so the design chat stays clean
• If you'd rather skip it this time, just say so

Filter My reviews by project

With many reviews, finding an older one by project was hard. The full list now shows which project each review belongs to, and you can pick a project to see only its reviews.

• Each row shows its project name
• Pick a project in the project filter to see only its reviews, or only the ones without a project
• Works together with the status tabs, search and sorting

The web coach tells you what went wrong

When the map or the current step couldn't read a document, it looked like an empty document, and on phones you couldn't see the result of going back to a previous version. Now one line says what happened and shows how to try again.

• If the map or the current step can't read a document, you see "Couldn't load" with [Load again]
• On phones, the result of going back to a previous version shows above the document
• On the English screen, the coach's notices, tool activity and approval cards are in English too

Get the new version even with the web coach left open

PurplePrint improves often. But if you kept a web coach tab open for days, you kept using the old screen. Now a small notice tells you when a new version is ready.

• Press [Update] on "A new version is ready" and the screen switches to the new version right away
• Text you were typing stays in the input. If the coach is answering, you see the rest after the update
• If you want to skip it for now, just close it. It won't ask again for that version

Long conversations now open right away in the web coach

When a design conversation got long, opening the project could take a few seconds, and the screen could stutter while the coach's answer streamed in. The latest part of the conversation now shows right away, and the earlier part fills in after.

• Conversations with over a thousand records open in under half a second
• The screen no longer stutters while the coach's answer streams in
• You land on the latest message when you open it

A review of separate material can now continue as a project

When you asked your AI tool to review some material right away, there was no way to tell which project that review belonged to. Now the coach asks once, when you get the result.

• If it belongs to a project you are working on, the review is attached to it and shows under that project in your review list
• If it is a new idea, the coach creates a project and starts the design from the material you reviewed
• If you only wanted the review, just leave it. The coach does not ask again for the same review

Your Hyper-Reviews are now grouped by project

As reviews piled up, it was hard to tell which project each one checked. Your review list is now grouped by project, and each group can be folded.

• Reviews run on a project's design documents sit under that project. The web list and the list in your AI tool match
• Reviews run on material you handed over separately sit under 'No project'
• Earlier reviews sit under 'No project' because we have no record of which project they came from

Screen design now makes one easy-to-miss button stand out for first-time users

Icon-only buttons keep a screen clean, but first-time users often don't know what they do and walk right past them. Screen design now picks one button that is easy to miss but valuable once used, and gives it a slow glow until it is pressed once. Conditions are set alongside it so the screen doesn't get noisy.

• Only one per screen, and once pressed it stops glowing on that device
• It pauses while a tour or another guide is showing, and stays still on devices with reduced motion turned on
• Data relationship diagrams in development design are written in a form that renders on the map

Current coaching version: v2.82.0.

Development design now places servers where your users are, and picks AI models by testing them

A slow service is often slow because of the network path, not the server code. The same hosting can route each country to a different data center, so requests from users in Korea can travel overseas first. So in development design, you now choose where the server and database live based on the country your main users are in, and measure it from that country before launch. If the product has AI features, models are not picked by name: you confirm each model is reached through a path that does not keep your inputs, then test them with the same inputs before choosing.

• No single AI company is recommended as the answer. Each feature picks a model by quality, speed, and cost
• AI features that handle personal data or text your users wrote only use paths that do not keep inputs
• Whenever a model or its thinking level changes, the same test runs again

Current coaching version: v2.81.0.

Delete finished projects from your archive

Archiving a finished project freed its spot, but it stayed in your archive. You can now delete archived projects.

• Say "delete [name]" for an archived project and it is removed from your PurplePrint list. The documents in your own folder stay as they are
• Projects in progress need to be archived first. Deleting cannot be undone
• Subscriptions allow 4 projects in progress and 8 archived. Deleting an archived project frees a spot

New PurplePrint connections use a faster address from Korea

Calling PurplePrint from your AI tool could mean a short wait each time, because the connection address went through an overseas route. New connections now use an address that goes through Seoul.

• The new connection address is app.purpleprint-ai.com/mcp. The connection guide, docs and install prompts all use it now
• Your current connection keeps working as is. You don't need to reconnect
• When you connect on a new computer or another AI tool, just follow the connection guide

Archive a finished project on the free trial to start a new one

During the trial you may want to set one idea aside and start another. You can now archive a finished project to make room for a new one.

• The free trial allows 2 projects in progress. Archiving one frees a spot
• You can archive up to 2 projects and bring an archived project back to continue it
• Projects deleted on the web also count toward the 2 archived

The web coach is much faster from Korea

Opening the web coach from Korea could take a long time, because the connection went through an overseas route. The web coach now opens at a new address that connects through Seoul.

• The web coach address is now purpleprint-app.pages.dev. The old address forwards you to the new one
• Sign in once more at the new address. Your projects and conversations carry over
• Using PurplePrint through an AI tool (MCP) has not changed

See your design as pictures, not just text, with core logic added to development design

A design written only as text is hard to take in at a glance. So the current way, the after picture, user journeys, screen flow, and data relationships are now also drawn as pictures inside the documents. In development design, before choosing technology, you first lay out how the service works inside (its core logic) as a picture and decision rules.

• Pictures are part of the document. When the text changes, the picture changes with it
• In the web coach, the Map in the progress tab shows documents and pictures at a glance, and you can pin spots to send to the coach
• When you use PurplePrint through an AI tool, it also makes a clickable mockup that follows the screen flow once wireframes are done

Current coaching version: v2.80.0.

Free trials now include 1 Hyper-Review, and Ultra opens on a higher plan

We adjusted the number of server reviews in the free trial so more people can try PurplePrint. You can still get a review of your own document during the trial.

• New free trials include 1 Hyper-Review. If your trial already started, you keep 3
• Ultra (a deeper review that keeps digging until no new findings appear) opens on an upcoming higher plan. Basic's 10 standard reviews a month stay the same
• Reviews that could not start because none of the attachments were readable do not count toward your monthly total

Design a demo and tour that show your whole product before sign-up

People who arrive for the first time do not create an account before they see what they get. So if your product starts with an account, the screen design step now plans a pre-sign-up demo and tour by default. The demo shows real results your product made, and the tour walks through every feature down to the parts inside each screen, then ends with one next action.

• The demo is read-only. Calls that cost money are hidden on screen and blocked on the server too
• When features change, updating the tour script is part of the development checklist

The current coaching version is v2.79.0.

Start, wrap up, continue: the three basics, explained

New and existing users alike often missed these three. The coach now mentions them once, briefly, when you start for the first time and when you save a document.

• Start: in a new chat, say "퍼플프린트 시작해줘" (Start PurplePrint) and the coach picks up where you left off
• Wrap up: when you are done for the day, say "다음 세션에서 이어가자" (Let's continue next session). The coach saves what you decided to your documents
• Continue: keep going in the same chat, or start a new chat when it gets long
• This applies when you use PurplePrint through your AI tool

After design, development follows the same way of working

When you keep building a fully designed project in your AI tool, each development session now receives PurplePrint's way of working. Before changing anything, it checks the design documents the work is based on, and after changing it, it checks again for anything missed. If the design itself has to change, it updates the documents before the code.

• Each time a piece of work is done, it takes care of the documents that should change with it, such as the task sheet and the handoff note
• When wrapping up a session, it briefly shows "What PurplePrint took care of this session", listing only things you can verify
• This applies when you use PurplePrint through your AI tool. The web coach stays the same

See what you have filled in at this step and what is left

Section titles in your design documents now start with that step's item number (for example, 4.2 Screen groups and navigation). Items that do not apply still get a section with a one-line reason, so a skipped item is never confused with a forgotten one. At the top of the web coach's Progress tab you see how many of this step's items are filled in, and clicking a filled item opens that section of the document.

• The progress bar now means this step's items everywhere. The full 10 steps are in the list below
• In the design gap check step you see each review round and how many issues it found
• For first-time builders, a short note in the conversation explains what each new step does and what happens if you skip it

The current coaching version is v2.78.0.

First-time design asks fewer repeat questions, and decisions land in your documents right away

For first-time builders, the coach used to ask one small question at a time. Now it sketches a whole flow and asks you to confirm it once. It asks separately only about things you must decide yourself, such as how money changes hands, and fills in the rest with what similar services do, marked as a default. It does not ask again about anything you already decided.

• Your decisions are written into the design documents during each phase, not only at the end, so they stay even if the conversation is interrupted
• The coach keeps one consistent, polite tone from start to finish

The current coaching version is v2.77.0.

Hyper-Review reports now come in your language

If you design in English, your Hyper-Review report now comes in English too: the server report, the report view, and the start and finish emails. Korean reports stay the same.

• The language follows the one you use with the coach. On the web it follows the text you submit, and starting from the English page always gives you English
• There is now an English sample report at /en/review/sample

From October 9, use the same design coach right in your browser, no install needed

Starting October 9 at 00:00 KST, you can use the PurplePrint design coach directly in your browser. Sign in and turn your idea into a design without installing or connecting an AI tool. The documents you make stay on the web.

• The trial covers the web coach through problem definition. After that, subscribe or bring the project into your own AI tool to continue
• Documents made on the web move into your AI tool's working folder in one step, so you pick up where you left off
• Connecting PurplePrint to your AI tool works the same as before

Get ready to be cited in search and AI answers before you build

If you plan to reach people through search or AI answers like ChatGPT, some groundwork can't be added after launch: recording when a page was last edited, telling apart visits that come from AI answers, and deciding how AI crawlers are treated. These now go into the pre-build checklist, and after launch the coach offers to review your search and AI visibility with you once a month.

• Nothing changes for services that don't rely on search or AI answers
• The coach now says things like "the people who will use this" instead of design jargon such as "persona"
• Already live? Ask the coach to "check my search and AI readiness" and it will start with what's missing

The current coaching version is v2.76.0.

Your design documents now record how they connect

Each design document the coach creates now starts with a short header that notes which earlier documents it builds on, so the document map and the upcoming app can draw how your documents connect. The coach also drops the long dash that makes AI writing easy to spot and uses periods and commas instead.

• Documents you already have get the header the next time they are edited. Nothing to do on your side

The current coaching version is v2.75.0.

Your coach now flags errors and metric drops in your live service first

Once a service is live, some problems never get reported by users. Your service now detects errors and anomalies like a sudden drop in sign-ups and files them as GitHub issues. When you start a session, the coach checks those signals and user feedback first and tells you what is most urgent. Fixes happen only after you confirm.

• The pre-build checklist now includes operational signals. The coach picks which ones fit your service
• We also fixed an issue where the coach could barely read user feedback issues
• Already live? Ask the coach to "add operational signals" and it will set up what you need first

The current coaching version is v2.74.0.

See design options as real HTML screens, then choose

The design step (Phase 5) has changed. The coach first recommends reference sites and a starting template, then builds two or three versions of your landing and core screens and opens them in your browser. Your design and screen plan follow the version you pick.

• If your product is announced on the web, the landing is part of the screen plan and stays consistent through review and marketing
• Already past the design step? Ask the coach to "show my landing and core screens in a few versions" to redo it the new way

The current coaching version is v2.73.0.

Start your product landing from one of three variants

A SaaS Landing Template is now in Kits. Compare the original, refined, and product-first variants at desktop and mobile widths, pick one, then fill each block from your earlier design outputs.

• Numbers and testimonials come only from real data. If you have none, drop that block
• Available during your free trial too

Kits now install with their original folder structure

Until now, an installed Kit was merged into a single .md file, so template screens (HTML) could not be opened. Now the files are saved under engines/kit-name/ in your project folder exactly as they are, and you can open the screens right away.

• Large Kits are received in several parts. The coach fetches them all and tells you once at the end
• If you installed a Kit earlier and it was saved as a single .md file, install it again

You can install Kits during your free trial

Until now, Kits could be installed only with a paid subscription. Now you can also install the templates and engines that plug straight into your design and build during your free trial.

• Ask your AI to show Kits to see the list, and the coach picks the one that fits your current step
• After a trial ends, subscribe to keep installing

Hyper-Review now reviews every document of a long design set in depth

Inputs over 50,000 characters are split into segments at document boundaries, reviewed segment by segment, and merged into one report. Before, a long set was reviewed in one pass and the later documents got only one or two findings. On the same design set, you now get about three times as many findings, at least nine per document.

• Review time stays about the same (usually 20–40 minutes), and it still counts as one review this month
• Segments still produce a single report. When the same problem spans several documents, it is grouped under a root-cause theme
• If a report section such as team capability or next actions comes back empty, we fill it once more. If it is still empty, the report says so

Hyper-Review now reads long design documents to the end: up to 130,000 characters at once

Until now, only the first 7,000 characters of each document went into a review, and we didn't tell you. You can now paste a whole design set and up to 130,000 characters are reviewed in one pass. Review time stays the same for long inputs (usually 20–40 minutes).

• Beyond 130,000 characters, the progress screen, completion email and coach response tell you how far the review read: pick the parts that matter and run again to have them fully covered
• Attached files and documents from GitHub or Notion follow the same rule

From September 25, 22:00 KST, Hyper-Review AI processing gains a fallback path: Zero Data Retention stays

From September 25, 22:00 KST, review AI processing via Vercel AI Gateway adds OpenAI as a fallback path. Requests are processed on Microsoft Azure as usual, and on OpenAI only when Azure cannot respond: so reviews keep running during an outage.

• Both paths are requested only under Zero Data Retention terms: the design content you send and the results are not retained or used to train models
• Sections 6, 7 and 8 of the Privacy Policy (sub-processors · international transfers) change from September 25, 22:00 KST · until then, processing runs only on Azure
• If you don't run Hyper-Review, your design content is not sent anywhere

Details: /en/privacy

Hyper-Review now finishes in about 20–40 minutes: and findings no longer get cut off midway

From September 25, 19:00 KST, reviews run on the new AI processing path, and a run that used to take about 1–2 hours now takes 20–40 minutes. We also fixed AI responses being cut off at the length limit mid-review, which could drop some findings.

• We still email you when it's done: very large inputs can take a little longer
• As announced, processing runs only under Zero Data Retention terms: the design content you send and the results are not retained or used for training

From September 25, 19:00 KST, Hyper-Review gains one more AI processing path: your design content is never retained

To improve review quality and speed, from September 25, 19:00 KST some review steps may be processed by AI models run on Microsoft Azure via Vercel AI Gateway. This path is requested only under Zero Data Retention terms, so the design content you send and the results are not retained and are not used to train models.

• Sections 6, 7 and 8 of the Privacy Policy (AI processing by server review engines · processors · international transfers) change from September 25, 19:00 KST
• Until then, processing runs only on Cloudflare Workers AI · text extraction from attachments stays on Cloudflare going forward
• If you don't run Hyper-Review, your design content is not sent anywhere

Details: /en/privacy

Hyper-Review reports now group key issues by root-cause theme

When there were many key issues, a long list made it hard to know where to start. Issues that share a root cause are now grouped into theme cards, so you can see at a glance what can be solved together.

• Problems that need different decisions are kept separate: closing one issue makes it clear what is done
• The structure map is grouped by theme too, and two-digit issue numbers no longer get cut off or overflow
• If an internal identifier slips into report text, it is shown as the visible issue number instead

Your builder level is now told on a single 9-step scale

The level grid could show 8 while the coach said "your latest stage is 4", which was confusing. The coach now speaks only in how many of the 9 steps are filled, and recent records use the same scale so you can compare them with your current level directly.

Search registration now covers Naver, Bing and Daum: not just Google

The pre-development checklist now includes Naver Search Advisor, Bing Webmaster Tools and Daum Webmaster Tools. If your users are in Korea, Naver and Daum matter: Google registration alone doesn't get you listed there.

• Bing can be registered in one step by importing from Google Search Console
• Already live? Ask the coach to "check search registration" and it will register the missing ones with you

Hyper-Review now reads attached PDFs and images: and tells you what it couldn't read

The step that reads attached PDFs and images was blocked by an external AI service, so files could silently drop out of a review. Attachments are now read in the same place as the review itself and are never sent to an external AI service.

• Large screenshots and diagrams are read in pieces, down to the small print: text that can't be read with confidence is marked rather than guessed
• If a file can't be read (like a scanned PDF) the progress screen and the completion email tell you why and what to do instead
• PDFs are read only when they contain text: for scans, upload the pages as images (PNG or JPG)

A Codex login that takes over 5 minutes no longer ends at a dead end

Codex only waits 5 minutes for a login. First-time users could take longer: their account was already created, yet the flow ended on a “127.0.0.1 refused to connect” screen. Now, if it runs late, you'll see “You're signed in: press login once more in Codex.” The second time you're already signed in, so it finishes quickly.

Hyper-Review is much faster: and reports come back complete

A review took 3–4 hours, and the team-capability diagnosis in the report often came back empty. The cause: long AI responses were cut off at about 4 minutes and had to start over. We now receive responses in pieces as they stream, which fixes both.

• A run now takes about 1–2 hours instead of 3–4: we email you when it's done
• The team-capability diagnosis (strengths, gaps, coach note) is always filled in
• More reviews run at the same time, and reviews that stop midway on our side are cleaned up automatically: they don't count toward your monthly reviews

Your builder level keeps growing after design, while you build

The builder level updated each time you finished a design document: so once you completed the design and moved on to building, there were no documents left to finish and your level stood still. It now keeps growing while you build.

• When you save progress or record a new deliverable, the coach looks back at the work you just did and updates your level: about once every few days, so it never breaks your flow
• Next to your level, you'll see the date of your latest record and the next step it left you
• Nothing is made up: if there's no record yet, that line stays empty

Your conversations and project content never reach the server: only the stage is kept. To opt out, tell the coach "turn off auto recording."

The coach points you to the document it just made: and offers to wrap up when the conversation gets long

Design documents only deepen your understanding if you actually read them, and it was easy to miss that one had been created. And when a conversation ran long, earlier context faded while the session kept going. The coach now takes care of both.

• When a document is created, the coach links it in the chat and, where the environment allows, opens it for you
• When the conversation gets long, the coach offers first: "Shall we wrap up here and continue in a fresh conversation?"
• On wrap-up, the coach checks your documents against each other, fixes any mismatch, then saves: nothing for you to do

In a new conversation, just say "continue PurplePrint" and you pick up where you left off. Everything you made stays on disk.

Your builder level now shows up every time you finish a document

Until now you had to ask for a growth measurement to see it. Now, each time you finish a design document, the coach looks back at the stretch you just completed and shows how much of the 3×3 grid (the same one as our logo) you've filled.

• Nothing to invoke: the nine-cell gauge updates on its own when you finish a document
• Cells don't fill by producing more documents. They light up as you lead the design yourself, and a lit cell never turns off
• When you complete a design or light a new cell, the coach points out which skill grew and your next step

Your conversations and project content never reach the server: only the stage is kept. To opt out, tell the coach "turn off auto recording."

Explore and install product engines proven through real work in Kits

PurplePrint now gathers the methods, templates, and operating systems we use and neutralize for reuse in one library.

• Filter Kits by operations, web build, UI system, growth, or audit use
• Open each Kit for its overview, fit, pipeline, and expected outcomes
• Result previews are currently hidden while we select examples that meet the public quality bar
• See aggregate detail views and successful package installs without personal identifiers
• Copy an install request and paste it into the AI agent you already use

Kits are a download benefit included with a paid subscription. Trial accounts can browse, but installation requires a subscription.

Start from what you already have, then design how to validate the market

PurplePrint coaching is now at v2.68.0. It starts from your real materials, checks that your environment can preserve the work, and closes the validation decisions that must exist before launch.

• If you have a plan, Notion page, or repository, the coach reads it first and only asks for material it cannot open
• If your chat surface cannot create files, it warns you about losing artifacts and shows the supported path before coaching begins
• Phase 7 now defines use, willingness-to-pay, and retention hypotheses plus one experiment with explicit success and failure criteria

The current coaching version is v2.68.0. Existing projects do not need a restart; revisit only the Phase that needs reinforcement.

Sign in with Google or GitHub: it's the same account either way

Until now each sign-in method created its own account, so coming back through the other one hid your work. If the email matches, they're now one account.

• Your projects, reviews and growth records show up whichever way you sign in
• Billing attaches to that same account: your sign-in method no longer splits it

For safety, only verified email addresses are linked. An unverified address never connects to an existing account.

Finished projects can be archived: the slot frees up, the record stays

There was no way to tidy up projects, so someone ended up reusing a finished project's slot for a completely different topic. Instead of deleting, you can now archive.

• Say "archive [name]" and it leaves the list, freeing one project slot
• Nothing is deleted: "restore it" brings it back exactly as it was
• "Rename it" works too, so a project that changed direction doesn't need a new slot

Your output files live in your own folder and are never touched by archiving.

Your next single step now has its own place on the growth page

After opening a place to see the curve comes a place to read it. Pick a measurement and what you actually did (and what to do next) appears beside it.

• Pick a measurement on the left and that record opens on the right: clicking a point on the curve does the same
• "Your next step" is no longer buried inside the assessment sentence
• Trend, history and detail each have their own pane, so you always know what you are looking at

Older measurements have no next step recorded. We leave it blank rather than invent one: measure again and it will be there.

Your growth curve is now on the web

Until now a growth measurement showed once inside the session and vanished. Watching the curve build is the whole point, and there was nowhere to watch it. Sign in and your records are gathered in one place.

• Lead-stage trend as a curve: from the second measurement you can see what changed
• The seven behavioral signals from your latest measurement, at a glance
• Which projects you worked on between each measurement

Your conversations never reach the server: only the tallies. Project attribution covers work done through PurplePrint.

Session hand-off notes always land in the same place + growth can be measured in depth

When a long conversation had to be cut short, the coach invented a different filename each time for the hand-off note: so the next session couldn't find it. The name and location are now fixed. Growth measurement can also go much deeper on request.

• Hand-off notes live at `docs/review/handoff-DATE-keyword.md`: the next session restores from that one file
• The growth table now shows which project each entry came from, so the curve lines up with what you remember
• Ask "look at my growth in detail" and it reads far more sessions and cites the evidence behind each call

The detailed pass costs more time and tokens. The default stays light: ask for depth only when you want it.

Parallel projects no longer bleed into each other + finished designs hand off to build

With several windows open on different projects, a save could land in the wrong one. And people who had finished their design still only got "want to continue?". Both are fixed.

• Each window records only to its own project: the other one no longer gets dragged along
• A project whose design is complete gets build hand-off guidance (pass the handoff and task docs to your dev agent) instead of a resume prompt
• Growth records tell you when it's worth measuring again: the curve appears from the second measurement

With two or more projects, the coach always tells you which project it wrote to.

Design isn't decided by feel: measure references, then use the numbers

In the design step (Phase 5) the coach could skip actually measuring reference sites and go by feel. The result works but looks flat. Now the step doesn't close without measured numbers.

• Open 2-3 reference sites, measure type scale, section spacing and key element sizes, and set the design from those numbers
• A helper agent (Codex) can do the measuring: you just take the numbers
• Adjectives like "modern, clean" alone count as incomplete. The review step (Phase 6) checks against the same bar
• Three more "made by AI" patterns get filtered: color removed without filling the gap, every section on the same rhythm, the signature card shaped like every other card

Projects already past the design step don't need a redo. Just fill in the numbers in your design system.

A helper agent takes the research: your coaching thread stays clean

When the coach researches competitors, personas and markets itself, tens of KB of raw material piles up in the thread, and the design decisions that follow get dragged along by it. Now a separate helper agent does the research and the coach only reads back the result.

• With a Codex subscription: the coach walks you through install and login, then runs the helper for you. No need to open the Codex app
• Without one (or not sure): the coach does the research itself. We don't push you to subscribe
• Model and research depth are preset: you don't need to know what to pick
• You don't wait around: it runs in the background, and the coach picks up the result when it's done

This applies to competitor, marketing, persona and GTM research. Journey simulation and code audit stay with the coach, since those need the design context in hand.

Tighter screen design: a structure you don't get lost in, and guidance for first-timers

Three defaults landed in the screen-design coaching (Phase 4~5). Until now the coach decided these case by case.

• Structure first, screens second: decide how screens group, what they're called, and how many taps deep they sit, then draw
• Walk your journeys back over that structure: wherever “what do I tap next?” stalls, that's the gap
• First-timer guidance by default: empty-state nudges, a first-run tour, inline hints, progress indicators
• Icon buttons carry an explanation: no more buttons nobody dares to press
• Blocks the “made by AI” look: purple gradient heroes, endless rounded cards, and friends now need a stated reason

Already mid-design? That's fine. If you're shaping screen flow, reinforce the structure once; if you're settling the visuals, just add the icon and guidance specs. The review stage picked up the same checks.

Hyper-Review on the web: no install, connect GitHub or Notion

You can now get the 39-axis review straight from the web, without installing anything in your editor. Sign in and it's free, 3 times during the trial.

• Paste: drop in your plan or design as-is
• Attach a file: design PDFs and wireframe images are read too
• Connect GitHub: pick a repo and it reviews the code and docs as they are
• Connect Notion: no need to move your planning pages anywhere

Connecting is one button, and only the repo or page you pick is read, read-only. Disconnect any time. The report states what was read and what was left out: because no findings on a file that was never read doesn't mean it's fine.

Hyper-Review series: from service design to code alignment

Ahead of the July 23 web launch, we finalized the three-layer review contract from service design → dev design → code:

• Hyper-Review: checks your service design across 39 axes, recursing until zero findings
• DAR (Deep Architecture Review): reviews your dev-design architecture across 8 rotating lenses
• Code Audit: checks whether the code matches the design

Hyper-Review and DAR run asynchronously on the server. Code Audit keeps raw code local and runs through your agent. Standard is the base pass; Ultra recurses more deeply.

4 ops commands: recover on your own when stuck

server-health (self-diagnosis)·manual (detailed guide)·oauth (account·trial status)·handoff (artifact package). The agent diagnoses “connected but not working” and tells you the next step.

Hyper-Scenario: user-journey simulation

Replays your design as the persona, step by step, so you see where users stall and drop off before you build. Pulls User Flow·Story Map·test seeds too. Runs on your AI agent's web: zero cost.

Hyper-Research family: competitor·marketing·persona·GTM

Fills in what you can't know from your head with real research: competitor strategy·marketing channels·persona validation·go-to-market. In the Phase flow, the coach proposes it at the right moment. Runs on your AI agent's web: zero cost, content stays local.

Connection stability: slash visibility + 30-day login

/purpleprint:* slash shows reliably in Claude Code·Cursor. One login lasts 30 days, so no re-auth every time.

2-week free trial + beta billing

Sign up and a 2-week free trial starts with no card. After that, beta $13/mo (list $20).

Builder-capability measurement: my-growth

Measures how much you've grown as a builder while shaping your design. Content-free: it never reads what you wrote, only behavioral signals (like the decisions you made on your own). It shows 'Zero to Builder' as measured evidence, not a promise.

PurplePrint MCP beta: first release

Phase 0~9 conversational design coaching inside your coding tool. From idea → a build-ready design.