Vizebel
Changelog

What’s new

Vizebel is in beta. Every release since launch, newest first: what we added, what we improved, and what we fixed.

Releases

Autopilot: the routine work on your content gets done for you

Staying on top of your content means going back to your Google data and pages, seeing what changed and working out what to do next. Switch on Autopilot and Viz does that on the schedule you set, runs the routine steps itself and emails you what got done and what needs a decision. Nothing is published.

New

  • Viz refreshes your Google data and pages and checks your site when they need it, with no one having to remember.
  • Each run lines up to five next steps with the reason for each, so you start from a short list of decisions.
  • One email after each run says what finished, what’s still going and what’s waiting for you. Approve costed steps from the email or from the new Today page, which keeps the record of each day.
  • Hand over more when you’re ready: page reviews, snapshot re-runs and question gathering can run without asking. Page rewrites and changes to your strategy always wait for you.
  • Set a monthly credit cap, which planning each run counts towards. If the cap or your credits run out, we email you and Viz stops lining up new steps until there’s room.
  • If a run couldn’t plan or a step didn’t finish, start it again from Today with Run again.
  • Choose every day, weekdays or Mondays, the time it runs and who gets the email. Available on paid plans.
  • Through the MCP, Claude can read the day’s steps, then approve or dismiss them.

Audiences by situation, Viz that takes action, and link checks

Each audience now describes the situation a customer is in and what they need to know before they decide. Ask Viz can now do things for you, and the Technical tab checks every link on your site.

New

  • Audiences are written around the situation a customer is in: who is usually in it, the jobs they need done, what they need to know before deciding, and the words they use. Your topic map starts from your focus areas, and audiences show who is asking in each one.
  • Add an audience from a description, your notes, a brief or a call transcript, and redraft one in place from new notes. Its topics and questions stay with it.
  • Who else might you be talking to? Vizebel suggests audiences your site doesn’t speak to yet, to add or dismiss.
  • Ask Viz can now act for you: review or rewrite a page, start a draft, gather customer questions, check AI answers, manage audiences and more. Anything that costs credits or removes something shows as a card with its price, and runs only when you tap it.
  • Link checks on the Technical tab: pages missing from your sitemap, pages nothing links to, links to pages that fail or redirect, and broken links to other sites. Each finding opens onto the page, the link text and the fix. Links are checked after each page refresh and every week, at no cost.
  • Vizebel now finds pages your sitemap leaves out by following the links on your site.
  • Each page’s Technical checks tab lists the links to and from it.
  • Through the MCP, Claude can add, redraft and remove audiences, settle suggested audiences and rebuild your topics.

Improved

  • Filter AI visibility questions by focus area and topic as well as audience.
  • Refreshing out-of-date Search Console data fetches only what’s new, so it finishes much sooner.

Fixed

  • A new workspace no longer re-imports its pages on its first night.
  • Gathering customer questions while a new workspace finishes setting up no longer gathers the same topics twice.
  • Fewer AI Overview checks and question gathers fail during a brief outage at our search data provider.

Use Vizebel from ChatGPT, Codex, VS Code and Cursor

Work in Vizebel from the AI app you already use. The MCP now connects ChatGPT, Codex, VS Code and Cursor as well as Claude.

New

  • Connect ChatGPT, Codex, VS Code or Cursor through the MCP, alongside Claude and Claude Code. Every app gets the same tools, works within the workspaces and permissions you approve, and spends within the monthly credit limit you set.
  • The MCP page has a tab for each app, showing what to paste into it: the connector URL, a command, or the entry for Cursor’s mcp.json.

Improved

  • The MCP help guide walks you through setting up each app.

Edit your audiences any time, in Vizebel or from Claude

Keep your audiences up to date as you learn more about your customers. Edit one straight from its card, or let Claude read and edit each audience in full through the MCP.

New

  • Edit an audience straight from its card, even after you’ve confirmed it, so your drafts and article plans are written for the customers you have now. Editing is free.
  • When an edit changes an audience’s situations, jobs, pain points or worries, Vizebel offers to refresh your topics, so your topic map keeps up with who you’re writing for. The credit cost shows on the button.
  • Through the MCP, Claude can read each audience’s full profile and, on a connection allowed to make changes, edit it for you. You approve each change.

Fixed

  • Pages in a new or updated sitemap now show up the next time you refresh your pages.

Home, AI visibility in three tabs, and visits from AI assistants

Each workspace now opens on Home, which shows where you stand on each of the five questions. AI visibility is rebuilt around an overview, and Analytics now shows the visits AI assistants send you.

New

  • Home: see where your workspace stands at a glance, with a card for each of the five questions, from who you’re talking to through to what to change next. Each opens on its figures and the place to act. Once Getting started is done, Home leads with the article worth writing first.
  • Visits from AI assistants, from Google Analytics: see how many visitors ChatGPT, Gemini, Perplexity, Copilot, Claude and other assistants send you, by assistant and by the page they land on. They show on Pages, AI visibility, any page or topic that had visits, and the monthly report. They are counted apart from organic visits, and they are a minimum, since an app that hides where a visit came from shows up as direct.
  • Mark a customer search as not relevant from a topic page, so the topic’s counts, health and article plans reflect only the searches that matter to you. It stays marked when searches are gathered again, and can be restored.
  • Through the MCP, Claude can read what each AI engine said to a question, word for word, with the sources it cited.
  • Generate questions lets you choose chat questions, search questions or both, to match how your customers ask.

Improved

  • AI visibility has three tabs: Overview, Questions and Snapshots. Overview gives you the whole picture on one screen: Viz’s read beside charts by buying stage, by AI assistant and by who is named in your place, with up to four next steps and each snapshot’s last and next run.
  • An open snapshot is one page: its read, then its questions, with each answer opening under its row. Share of voice is measured against your listed competitors.
  • A topic page is easier to read at a glance. Its health sits between your pages and your reach, and each signal has its score, its change since last month and a reading in its own units. Customer searches are grouped under the page that answers them, so you can see what each page covers.
  • Keywords and Content tell you when Search Console or Analytics data is out of date, with a button to refresh it.
  • A snapshot that doesn’t re-run monthly and last ran over 30 days ago is marked Due a re-run.
  • Studio steps say what they’re doing. A fact that isn’t in your interview answers can be kept, removed or edited, and one regenerate applies every choice.
  • After you publish a page rewrite, the check tells you whether the draft is on your live page yet, in full or in part.
  • Refresh topics and Re-plan ask first, and tell you what stays, what is replaced and what it costs.
  • One name for each action, so buttons read the same everywhere: Refresh pages, Ideas, and Match topics to focus areas.

Fixed

  • A broken or hidden page no longer counts as answering a topic’s searches.
  • Viz no longer suggests a local search in a city outside your market.

Copilot joins the engines

AI checks can now ask Microsoft Copilot. Pick up to four of the five engines for each workspace in Settings.

New

  • Copilot as an AI engine, alongside ChatGPT, Gemini, Claude and Google AI Overview.
  • New workspaces start with ChatGPT, Gemini, Copilot and Google AI Overview. Existing workspaces keep the engines they had.

Send your monthly report as a private link

Share this month’s report with the people who need it, without attaching a file. Only they can open it, you can see when they have, and the report now has room for your own take on the month.

New

  • Share the report from the Reports tab. Type in who it’s for and we email them an invitation. They open it with a code sent to their inbox, so it stays private even if the email gets forwarded.
  • The Share panel shows who has opened the report and when. From there you can add or remove people, resend an invitation, change when the link expires or switch it off.
  • Someone you’ve shared with can open the report to their colleagues at the same company, and we email you when they do. Personal addresses like Gmail can’t be used to open it up this way.
  • Write your own commentary straight into the report: click under any heading and type. Or press Draft commentary and AI writes a first pass for every section from the month’s figures, ready for you to edit. It costs 1 credit and saves as soon as it’s written.
  • Each snapshot’s overview shows how often Google answered your questions with an AI Overview.
  • Admins can rename their organisation in Account settings.

Improved

  • On a phone, the report lays each topic out as its own card, so it’s easy to follow on a small screen.
  • Your agency’s name now sits beside its logo on the report cover.
  • Inside a client’s workspace, your agency shows in a tab above it, so it’s always clear whose account you’re in.
  • Sites that send visitors to a local version, such as /us or /au, are read as the version for your market, the way Google and AI assistants see them. Test bot access also tells you if assistants are being sent to the wrong country’s page.
  • Reconnecting Claude remembers which workspaces it could already reach.
  • A connected Claude no longer shows a badge. Only apps we don’t recognise are marked.
  • A rewrite always shows when it was drafted, so a fresh draft is easy to tell from the last one.

Fixed

  • Uploading your organisation’s logo now works for every admin.

The reader test, and a Rewrite tab in three steps

Every Studio draft and page rewrite now goes through the reader test before you publish. The Rewrite tab is rebuilt as three steps, with one list of what to fix.

New

  • The reader test: a separate AI model reads each draft as the customer it is for and scores whether they would read it, trust it and act on it. It marks the lines that would lose a reader and says why, and lists any name or figure that is not in your answers or your page.
  • Studio's final step and every page rewrite run the reader test. A draft that falls short gets one more pass at what the test listed, and you see the better of the two.
  • Fix a rewrite from its reader test: rewrite a line, edit it yourself or improve the whole draft, then check again. None of this costs credits, up to 8 checks or improvements and 30 line fixes on each draft.

Improved

  • The Rewrite tab is three steps: what the draft is written from, the draft, and ready to publish. Your facts and answers fold to one line once the draft uses them all.
  • One To fix list under the draft: names and figures not in your facts, the rewrite's notes, the lines the reader test marked and how the article is laid out, each with Rewrite it or Edit it. Click a quote to see it in the article.
  • Change an answer or a fact check after the draft is written and the tab lists what changed. Update the draft rewrites it from your latest answers.
  • Ready to publish shows the draft's score beside the live page's, and four checks: reads well for your customer, matches your facts, written from your latest answers and no examples missing.
  • A band at the top of the tab says where the draft stands, such as how many things are left to fix.
  • In the page review, "Reads as written by a person" is now "Reads naturally".
  • Find focus areas, Match topics, Refresh topics, Analyse citations and Review this page show their credit cost on the button.

Articles to write, and rewrites that start from your facts

Vizebel now turns what customers ask, what AI answers say and what your pages miss into a short list of articles to write. Page reviews are rebuilt on one writing guide, and a rewrite starts by checking your facts and asking for your own examples.

New

  • Articles to write: related ideas, customer searches, keywords and AI-answer questions become planned articles, each with its audience, its sections, the evidence behind it and why it ranks where it does. Where you already have a page, the plan is to improve it. Each topic keeps three open at a time.
  • Develop in Studio carries the plan into the draft, with a must-cover list of its searches and questions. The draft shows which it covers, mentions or still misses, and the final draft lists the claims to check before you publish.
  • A rewrite starts with a fact check. The review lists the statements a reader would act on that could be wrong, each beside its sentence on your page. Mark each one fact-checked, edit it or remove it, and the rewrite follows your answer.
  • The review asks for up to five first-hand examples the page lacks, and the rewrite puts each answer where it belongs.
  • Review a finished draft on the same checks as the page, for a score before you publish. Once it is live, Vizebel reads the page again and says how much of the draft it carries.
  • Pages that read as one: location, service and category pages that differ by a few swapped words are grouped, with the fix for that kind of page. A "last updated" date that moved without the text changing is flagged too.

Improved

  • Page reviews score two things: whether the page answers the question it is for, and whether it is original and trustworthy. FAQ blocks, summary boxes and FAQ markup no longer earn or lose points, and "Reads as written by a person" checks every rule in our voice guide.
  • Older reviews say they were scored on earlier checks, so a score never changes without a new review.
  • A rewrite sits beside your page, keeps its images and videos, and marks anything still missing as a gap to answer or leave out. Copy it as HTML or Markdown once no gaps are left.
  • A page opens on tabs: Overview, Content review, Rewrite and Technical checks. An article's technical checks now look for a named author with a byline and profile link.
  • Content, Technical, Keywords, Ideas, Studio and each snapshot open on tabs, with an overview before the long list. Content gains an Issues tab for fixes shared across your pages.
  • Ideas splits into Articles to write, AI answers, Google searches, Get listed and Questions to track.
  • Viz reads a page's position only once Google has shown it often enough to read.

Fixed

  • The page review names an article's author, not the raw markup some sites publish.
  • FAQ questions held in an accordion are kept when a page is read.

Work in your workspace from Claude

Connect Claude to Vizebel and work in your workspace from a conversation: read your topics, pages and Google data, check AI answers, review pages and draft articles. Credits are now taken only once the work can start.

New

  • Claude access, on the Max and Agencies plans. An admin turns it on under Account, then you connect Claude and choose the workspaces it can reach and whether it may read, make changes or spend credits.
  • From Claude you can read audiences, focus areas, topics, gaps, pages, keywords, Search Console and Analytics figures and AI answers; review and rewrite pages; gather what customers ask; and open and work through Studio drafts.
  • Anything that costs credits is quoted first and runs only once you agree to that price. Each connection has a monthly credit limit you can change on Claude access, beside its spend this month and its activity.
  • Save an AI-answer check from Claude as a snapshot, re-run it later, or switch it to re-run monthly once you agree to the monthly price.
  • Ask Claude to take a keyword, question, topic, focus area or a page off a topic. You approve each one, and a removed topic or focus area is archived, not deleted.
  • Where you serve customers in person: Vizebel reads the towns, cities or regions from your site, and you can edit them in setup or on your profile. Suggested keywords name only those places, and a business not tied to a place gets no local terms.
  • When ChatGPT ends an answer by asking you to narrow it down, such as by team size or budget, its follow-up questions and their options show under the answer.

Improved

  • Credits are taken only once the work can start. A page review, technical check or rewrite that fails is refunded, and the refund shows in Usage.
  • A change in health compares only the signals measured both times, on topics, focus areas and audiences.
  • Getting started counts only the pages that load, so its page count matches the stage beside it.
  • Studio lists each draft under its own title.
  • Start free goes straight to setup, which reads your site for you. The separate free check at vizebel.com/check has closed.

Where each workspace starts from, and what changed since last month

Vizebel reads where your site is starting from, in its content and its Google data, and works from there. Each topic charts its health by month, the monthly report compares AI answers on the same questions, and a snapshot can re-run itself every month.

New

  • Where you're starting from: your content (no articles yet, starting out, slowed down or established), and Search Console and Analytics each from waiting for data through building history to established. Getting started shows the stages, then a card with one suggested move takes over until each is established.
  • Health by month on each topic page, with the change since last month beside the score and beside each signal.
  • The monthly report compares AI answers with last month on the same questions: how often you were named then and now, and the questions you won and lost.
  • Re-run monthly: switch it on for a snapshot and it re-runs once a month, at the credits shown on the button. A month your credits do not cover is skipped, and you are emailed.
  • When your site cannot be read, or carries little text, setup asks you to describe your business and drafts your profile, audiences and focus areas from your own words.
  • A Searches row in Getting started gathers what customers search and ask for topics that have none yet.

Improved

  • Focus area health is not scored until your site has articles.
  • Keywords read Not measured until Search Console has data.
  • Publish dates are read from structured data on every page, and from the date beside the headline when nothing else gives one.
  • A snapshot's Visible, Mentioned and Cited cards say how many questions they count, and the monthly report's engine bars say how many answers.
  • On Snapshots, the change in answers mentioning you compares runs of the same snapshot.

What customers ask, and Viz beside every chart

Vizebel now gathers what customers search and ask about each topic, and reads every topic in two halves: your pages and your reach. Viz sits beside the charts on every screen, and the app moves into one sidebar.

New

  • What customers ask: for each topic, Vizebel gathers the searches and questions customers ask from Search Console, People also ask, Google autocomplete, related searches, YouTube and Reddit, and matches each one to the page that answers it. New workspaces gather them during setup.
  • Health reads in two halves, your pages and your reach, on one ring split down the middle. A new signal, Answers what's asked, counts how much of what customers ask your pages answer, so a topic's score can move once its searches are gathered.
  • The topic page is rebuilt around the two halves: the charts down the left and Viz down the right, each note level with the chart it reads. A question wheel shows what customers ask, with the questions no page answers yet beside it.
  • Add a site that AI answers cite as a competitor, straight from the topic page.
  • Viz reads beside the charts on Content, Keywords, Tech, Audiences, each page and each snapshot. The topics and pages it names open with a click.

Improved

  • One sidebar holds every page, grouped, with Focus areas first. It folds to a rail of icons, and on a phone the menu fills the screen.
  • Every page opens on the same slim header bar: the title on the left, its actions and Ask Viz on the right.
  • Opening Viz moves its note into the drawer, and a question asked there carries on from everything the note said.
  • Audiences shows one card per audience, read in your pages and your reach. View audience opens the full profile, and you can edit it there.
  • Focus areas sets Viz beside the health chart, with the focus area cards beneath it.
  • Every health ring shows one score out of 100 for whatever has been measured.
  • A finished snapshot's AI answers count toward your topics the same day.
  • Keywords splits a long list into pages, and Ideas keeps its board and filters in the header.

Fixed

  • Admins of a direct workspace can reach Team from the sidebar.

How each audience is served, and a clearer read on Focus areas

Audiences opens with a side-by-side view of how well each customer is served. Viz's note on Focus areas reads your whole strategy and names the next moves.

New

  • Audiences opens with a dashboard comparing your audiences: how many of their topics are covered, how often AI answers name you at each buying stage, and the search clicks their pages earn.
  • Articles that serve no audience are listed under the dashboard, least-read first.

Improved

  • Viz's note on Focus areas reads the whole strategy: the signal holding it back, then the next moves for Google search and AI answers, and what is already working.
  • Each move says where to start and links to the page to edit. A move across several topics opens them together on Topics.
  • The Claude engine in your visibility checks asks Sonnet 5.5.

Fixed

  • Workspaces set up before the new setup flow are no longer asked to confirm their audiences again.

A quicker start, with your first visibility check included

Setup reads as a profile of your business and ends with your audiences. Your workspace then opens with a short Getting started list, and its first visibility check is free.

New

  • Your first visibility check, included: Vizebel reviews your key pages and checks your customer questions in AI answers, and the results fill in as they land.
  • A Getting started list at the top of Focus areas, with the next step always one click away.
  • Viz sums up Focus areas in a few lines, each linked to its evidence, and the conversation carries on from the note.
  • Choose which AI engines a workspace asks, in Settings. A run costs only the engines it uses.

Improved

  • Setup reads as a profile of your business, then asks what you want to be known for and who you are talking to. Audiences can be confirmed now or later.
  • Audience portraits drafted during setup use no credits.
  • Unspent free credits stay with you when you upgrade.
  • Dashboards sit on their own background, apart from the pages and topics below them.
  • Before Google is connected, Content, Technical and Keywords show one Connect Google prompt instead of empty charts. You can skip Google during Getting started and connect it later.
  • Each topic with no content behind it says why, from the pages read on your site.
  • On a phone, the menu sits at the top as a compact bar, and rows and tables reflow to fit.

Fixed

  • Switching between Pro and Max in the billing portal changes your plan, and an upgrade gives you the new allowance straight away.
  • A re-test on one engine is charged for that engine alone.

Faster on large sites, and page clicks that match Google

Every screen opens faster on a site with a lot of Search Console history, a page's clicks match Google's own count, and pages are matched to topics by how they stand out on your own site.

Improved

  • Topics, Focus areas, Keywords, Questions and a page's own screen open faster, most of all on large sites.
  • Adding a topic returns straight away, and its pages and keywords fill in when they are ready.
  • A page is matched to a topic by how far it stands out from the rest of your own site, so a broad site and a narrow one both get a fair match.
  • A page's traffic counts once, under the topic it matches best, so topic totals add up to your site.
  • Questions written from a topic always name the product or category, so they make sense asked on their own.
  • On a page's own screen, the date range sits with the cards it changes.

Fixed

  • A page's clicks include the ones Google does not name a search for, so they match Search Console's total for that page.
  • A page's month-by-month trend shows one point for each whole month.
  • A topic added inside a focus area lands in that focus area.

Focus areas in your own order

Drag your focus areas into the order that matters to your business, and the rest of the app follows it.

New

  • Drag focus areas into priority order. Topics, the trend chart and Ask Viz list them the same way.
  • Topics shows the focus areas no topic serves yet, with Add a topic in the card.
  • A Not in a topic filter on Content, for the pages no topic claims.

Improved

  • Rank is your Search Console position on every screen, and the ranking page is the one Google showed most for the term.
  • Content shows its page list first, and each card fills in as its numbers arrive.
  • Audience portraits reflect the mix of people in your market.
  • The app retries on its own through a brief interruption, instead of showing an error.

Fixed

  • Keywords with Search Console clicks no longer show as unranked on a topic.
  • Long drafts, such as a full topic map, no longer stop part-way.
  • Old Performance links open Content.

Choose the pages that serve a topic

Add or remove a topic's pages yourself, and coverage now counts every page that serves the topic, not only the best one.

New

  • Add a page to a topic by hand, or remove one, and the choice holds through every later match.
  • Each page on a topic shows its review result and how strongly it matches, with one button to review every unreviewed page.
  • A technical check for sitemap addresses that redirect.

Improved

  • A topic's coverage adds up across every page that serves it, and moves as soon as a review lands.
  • An address that redirects is no longer listed as a page of its own, so the same page is not counted twice.
  • Questions and ideas on a topic are plain rows, with a Test button on each question.

Fixed

  • A large Search Console history imports in full, and an interrupted import resumes within days of where it stopped.
  • A topic's covered, partial or gap status stays in step with its coverage.

Add another workspace under the same sign-in

Running more than one brand? Add another workspace from the same account, each billed on its own plan.

New

  • A Workspaces page: add another brand under your existing sign-in, up to 20, and switch between them from the menu.

A page's own year-on-year trend

Any page's numbers now compare against the period before them, not against themselves, and reach back over a year.

New

  • Preset date ranges (7, 28, 90 or 180 days) on Performance and on every page's own detail screen.
  • A page's own Search Console trend, month by month, reaching back over a year.

Improved

  • A page's clicks and visits now compare against the real period before them, not a split of the same window.
  • Content pages load faster.

Focus areas

Getting Started now settles the broad spaces you want to be known for first, and Topics is built underneath them.

New

  • Focus areas: the broad spaces you want customers to know and choose you for, suggested from your own website and fully editable.
  • A Focus areas workspace showing health and movement from the topics underneath.
  • Topics grouped by focus area, with customer needs outside the chosen positioning shown separately.

Improved

  • Getting Started now asks you to settle your focus areas before Vizebel builds the topic map.
  • The workspace opens on strategy first, with Topics as the work underneath it.
  • The workspace prompts you to connect Google as soon as it's needed, instead of holding up the rest of setup.

Fixed

  • Organisation names are extracted correctly, instead of picking up part of a slug.

Co-branding, not white label

An agency's own logo sits beside the workspace it brought you; the header itself always stays Vizebel.

Improved

  • An agency's logo shows beside the workspace it belongs to, on its own tinted bar.
  • The header no longer flashes on load while it decides whose logo to show.

Your workspace's pulse in the header, and a rebuilt public site

A live indicator follows you around the app, and the public site now walks the same five questions the product does.

New

  • A live indicator in the header shows what's running or needs your attention, on every screen.
  • The arrival strip on a new workspace updates live and clears itself once the work lands, instead of sitting there unchanged.
  • A rebuilt public site: a new homepage, one product page covering every feature, and a reworked free check that hands off to signup.

Improved

  • Sign-up now ends at Topics, instead of waiting through every setup step first.
  • Logos display at the size they were given, instead of being boxed and shrunk to fit a square.
  • Content and Technical load only what they show, so large sites open faster.
  • Getting Started drafts fifteen to twenty starter questions across every priority topic.

Fixed

  • A Google sync interrupted mid-run resumes where it left off, instead of losing history silently.
  • Only one Google sync runs per workspace at a time.

A fuller public API

The public API returns the same measured results the app counts, and new workspaces get a welcome email.

New

  • The public API says whether each result was measured, and why not when it was not, so your own rates match the app.
  • Google AI Overview results in the public API include videos, citation pills, related links and ads.
  • Runs in the public API include their market and country.
  • New workspaces get a welcome email.

Fixed

  • A mistyped vizebel.com address shows a Page not found page.

Measured, not assumed

AI results say when an engine did not answer, every technical finding shows what was seen, and Search Console reports which pages Google holds.

New

  • Search Console tells you which pages Google has indexed, the address it chose, and when it last crawled. Inspection runs daily.
  • Every technical finding records what was observed: when the page was read, the status, the redirects, and the line or header quoted.
  • Checks for X-Robots-Tag and snippet controls, for the site and for each page. A page marked nosnippet cannot feed AI Overviews.
  • Crawler access is checked per group: search, live fetch, and training.
  • Every question says whether it is written for chat or for search.
  • Crawler checks read Cloudflare’s AI crawler settings and name them as Cloudflare’s, not yours.
  • Access checks also try the fetchers ChatGPT, Claude and Perplexity use when a person asks them to read a page.

Improved

  • AI results tell a brand that was absent apart from an engine that did not answer. Rates count only what was measured.
  • robots.txt is read the way Google reads it, with Allow, wildcards and end anchors.
  • A missing llms.txt is a note, and missing schema markup is a warning, not a failure.
  • A Core Web Vitals reading that fails says why.
  • Answers highlight the exact words that matched your brand.
  • A refresh you leave part-way is still running, with its progress, when you come back.
  • Fetched page text is read as content, never as instructions, in every AI prompt.
  • A check with nothing to fix shows a tick.
  • ChatGPT searches the web only when it chooses to, as it does for a real user, so rates match what people see.
  • The free check asks ChatGPT from the country your site shows, not only the one its domain suggests.
  • Visits on Content are labelled as organic search, and Google connections show when each source last synced.
  • The public site loads faster.
  • Articles now live at vizebel.com/blog. Old links still work.

Fixed

  • Help centre pages no longer appear under Articles.
  • A site whose address was corrected during setup is fetched from the corrected address.
  • Schema markup counts are reported against the right number of pages.

A more accurate read of your site

The crawler now reads a page the way a browser does, so titles, descriptions and page types on Content match what a visitor sees.

Improved

  • Page titles, descriptions and page types are read more reliably.
  • Title advice is specific to whether a title is too short or too long.
  • The free check lists what it will find before you run it.

Fixed

  • Topic priorities stay in step when you move several quickly.

Free credits, a help centre and your own logo

Ways to earn credits, public help pages, and white-label chrome for agencies.

New

  • A Free credits page: introduce yourself, report a bug, or link to us.
  • A public help centre and support page at vizebel.com/help and /support.
  • Agencies can show their own logo in the app header.
  • An affiliate programme at vizebel.com/affiliates.

Improved

  • Audience details can be edited directly, without regenerating the audience.
  • Every delete asks you to confirm first.
  • Sign-up from the free check carries your website through.

The free check asks ChatGPT about you

vizebel.com/check now writes three buyer questions for your site, grades your articles against them, and tests one in ChatGPT.

New

  • Three buyer questions written for your site, and up to four articles graded against them.
  • Pick a question and see who ChatGPT names and what it cites.

Improved

  • The public site was rewritten in a plainer voice, with a new colour scheme.
  • Sites that moved between www and the bare domain are read correctly.

Traffic on Content, and report history

Search clicks and visits sit above your page list, and every report you create is kept.

New

  • Content shows search clicks, visits and content score as trends.
  • Every report is kept, so you can open any past period.
  • Exclude a host or path from your traffic numbers.
  • Technical findings say what to do, not only what was found.

Improved

  • Rank comes from Search Console wherever it is connected.
  • One menu from the header on every screen, at every width.
  • Ask Viz knows the current Topics, Tech, Content and Report screens.

Tech and Content as their own tabs

The technical check was rebuilt around a full list of findings, and a workspace now opens on Topics.

New

  • Tech is its own tab, with stat cards and a trend over every finding.
  • Ignore a finding and it stays silenced until it gets worse.
  • Site-wide findings name the pages they affect.

Improved

  • A workspace opens on Topics.
  • Page detail leads with a preview, stat cards, and one list of fixes.
  • Check site re-reads your site before it audits, so findings are fresh.

Better questions, and snapshots on one page

Question generation was rebuilt, each question has its own page, and every idea knows the page shape that wins it.

New

  • Questions come in two registers: conversational, and the search form Google AI Overview runs.
  • Every question has its own page with the answers, who was cited and content opportunities.
  • Add a condition to any question and see whether the answer changes.
  • Ideas carry their buying moment and the page shape that wins it, and Studio writes that shape.

Improved

  • A snapshot is one page: questions, answers and trend.
  • Keywords shows 90-day trends, split branded from non-branded.
  • The app moved to soft-shadow cards and full-width pages.

Ask Viz, competitor briefs and a clearer Keywords page

The assistant has a name, a citation can become a content brief, and every keyword reads as a sentence.

New

  • Analyse citations turns the pages an engine cited into a brief for your own page.
  • Every page on your site has its own page in the app, with its traffic and checks.
  • Ask Viz remembers conversations and reads your Search Console data.
  • Add a topic by typing it, or describe it and let us draft it.

Improved

  • Ask AI is now Ask Viz, in the header on every screen.
  • Topics is a board of five priority cards, each with a health ring and next step.
  • Keywords reads as sentences: rank, landing page, clicks and where the phrase appears.
  • Subdomains count as your site everywhere.

UpliftAEO is now Vizebel

Same product, new name and domain. The app now lives at vizebel.com, and Getting Started leads with topics.

New

  • Getting Started drafts audiences, then topics, then questions, each waiting for your confirmation.
  • Choose five priority topics by dragging them into slots, or write your own.
  • Connect Search Console and Google Analytics during setup.

Improved

  • A new logo and typography.
  • Starter questions are written from your priority topics, in the audience’s voice.
  • Getting Started shows real progress with a page count.

Topics, Keywords, Google connections and a monthly report

The biggest release since launch: a map of the topics your buyers ask about, keyword tracking with Search Console, and a report to send.

New

  • Topics: a map of what buyers ask, scored on how well your site covers each one.
  • Keywords: add or draft terms, check rank, and see which pages hold them.
  • Connect Search Console and Google Analytics with a Google sign-in.
  • A monthly report with findings and deltas.
  • Core Web Vitals in the technical check, in plain language.
  • Audiences are persona cards with a portrait, bio and quote.

Improved

  • Navigation is a tab bar with a Topics health ring in the middle.
  • Results merged into Questions, with inline answers and status chips.

Faster runs

All four engines run at once, and results appear as each one answers.

Improved

  • A run finishes three to four times faster.
  • Results appear live during a run.
  • Test several questions at once.
  • Agencies see a credit breakdown by action.

Citations inline in answers

Answers show where each claim came from, and you can see how a crawler reads any page.

New

  • Citation markers sit inline in each engine’s answer.
  • Google AI Overview answers show their video cards, citations and related links.
  • See how a crawler reads any page, from Content.
  • Articles at vizebel.com/blog.

Improved

  • Content marks broken and missing pages instead of scoring them.
  • The free check samples several pages and finishes in about 20 seconds.

An assistant in the app

Ask AI answers from your own data and offers the next action.

New

  • Ask AI, an assistant grounded in your workspace data.
  • A searchable help page, with a guide beside every section.
  • Contact support from inside the app.

A free site check

New

  • A free site check at vizebel.com/check, no account needed.
  • Findings from the check carry into your new workspace.
  • Test whether each AI crawler can reach your site, on Technical.

Improved

  • Credits can be bought with or without a plan, and never expire.

Google AI Overview joins the engines

New

  • Google AI Overview is the fourth engine.
  • Rerun a single engine on a question.
  • Most cited brands and Top cited pages on Questions.
  • Topic ideas from People Also Ask land on the Ideas board.

Improved

  • Every column on Questions can be sorted.

API access, simpler sign-up and mobile support

New

  • Read-only API keys per workspace, with an export API.
  • Agencies can subscribe self-serve.
  • Hide pages in Content, with support and utility pages excluded automatically.

Improved

  • Sign-up asks one question: single workspace or agency.
  • The app works on a phone.
  • Question generation avoids duplicates and follows your market’s spelling.

Vizebel launches

The first public release, under the UpliftAEO name: the public site, the app, and paid plans.

New

  • Snapshots: freeze a set of questions and run them each quarter to track the trend.
  • A Technical audit with a PDF download.
  • Sign up as an agency or a single workspace.