Changelog
What changed in brandflare, newest first. Plain English, no version numbers.
Brands and leaders get their own tabs on the dashboard, once you track both
- If your workspace tracks both brands and a leader, the dashboard now splits into "Brands" and "Leaders" tabs instead of mixing them into one grid. Tracking only one kind still shows a single plain grid, same as before.
- Each grid now ends with an "Add a brand" or "Add a leader" tile, so adding another one no longer means leaving the dashboard to find the button. The sidebar's brand list splits into matching "Brands" and "Leaders" groups the same way.
A check you are waiting on no longer waits twice on an app that is not answering
- When one of the engines' own apps does not answer a check at all, a brand's first check and Check now used to wait a quarter of an hour, send the same questions again, and wait a quarter of an hour more before asking that engine another way. They now ask the other way after the first wait. Scheduled checks still give the app a second chance, since its answers are the ones people are shown.
Switching to Plus when you cancel now takes effect at renewal, as it says
- The "Switch to Plus" offer in the cancel dialog used to change your price right away: a Starter workspace got Plus free until its next renewal, and a Pro or Agency workspace on a yearly plan was moved down to Plus immediately, losing brands, seats and markets for the rest of the year it had paid for. It now does what the dialog always said: the price changes only when the period you have already paid for ends.
- Emails are retried if they fail to send, including the warning before a trial's first charge. A failed send used to look the same as a sent one and was never tried again.
- We've gone through security and performance across the app, following a weekly audit, and fixed what needed fixing: a staging sign-in gap, a key-in-a-link risk on the API, an unsubscribe link that never expired, and the slow spots it found. Nothing changes in how you use brandflare.
Agency's one-off reports are no longer free
- Agency no longer includes two free one-off reports a month. Every one-off report, on a brand you don't track, is now $99 each from the first.
A leader's scorecard gets an AI-written summary, and questions are sorted by type instead of topic
- A tracked leader's Overview now opens with a short AI-written summary of the scorecard: where they come across well, where they don't, and anything genuinely notable, written fresh after each check rather than shown live on every visit. It carries a small note that AI can get things wrong, and links the categories it calls out by name straight to the matching questions.
- On a leader's scorecard, "Clean" is now called "Trust": the share of a category's answers with no open accuracy or completeness flag. The By engine and By category tables both show Strongest and Weakest, scored on tone and trust respectively, so the wording and the two tables now agree.
- Questions & Prompts no longer filters by topic. Every question, including ones you add yourself, now has a single type instead (the kind of question it is, not its subject), shown as a filter pill and a tag on the question, with Other for the rare one that fits none of the usual kinds.
Tracking a named executive, mentioned on the public pages
- The homepage, How it works and Pricing now say plainly that you can track a named executive, founder or spokesperson the same way you track a brand: their name, title and organisation, no website needed. Nothing about the feature itself changed, only that it's now mentioned where you'd look for it.
A new brand's first questions are 8 branded, 32 unbranded
- Adding a brand on Trial, Plus, Pro or Agency now writes a fixed 40 questions straight away: 8 that name the brand directly and 32 that don't, in place of a model-judged set of 12 to 18. Most of your visibility only means something when an engine mentions you unprompted, so starting with that split keeps the brand's first read from being skewed by questions that already handed the engine its name. You can still edit, add or pause any question afterwards. Starter, whose room holds 15 questions, is unchanged.
API keys say who made them, and a round of security fixes
- The API keys list in Settings now says who made each key. When someone is removed from the team, the keys they made stop working at once, so a key copied before they left can't go on reading your brands.
- If a payment fails, your plan keeps working while it is retried, but new markets, one-off reports and knowledge base sources can't be added until it is paid. If every retry fails, the workspace moves to Free and keeps everything it has measured.
- A free trial now starts with a card. Other ways of paying can be used once the plan is paid for.
- Brand names can be up to 120 characters long, and research for a brand can be retried three times a day.
- A workspace can make 300 API and assistant requests a minute in all. Each key, and each person's assistant, can still make 60.
- An AI assistant connected by signing in now has to have been signed in for brandflare itself, so an assistant connected before today asks you to sign in once more. A key can no longer be put in the connector's address: give the assistant the address alone and sign in, or send the key in a header.
- The old client report links (brandflare.io/r/…) no longer open. They could not be switched off once shared. To share results with a client, use Export Report in the brand's menu.
- We've tightened how sign-ins, links in emails, reports and payments are checked, following a security review. Nothing changes in how you use brandflare.
Your results are yours, in writing, the privacy policy names every AI provider, a workspace can be limited to the providers' developer services, and an owner can delete their workspace
- Settings has a new card, AI providers. An owner or admin can limit the workspace to the developer services of OpenAI, Google, Perplexity and Anthropic, for organisations that allow only those. From the next check, ChatGPT, Gemini and Google Search are asked through OpenAI's and Google's developer services instead of their apps, so their answers may differ a little from what people see; Copilot, AI Overviews and DeepSeek stop being measured; and the daily visibility check asks Perplexity only. It is off unless you turn it on, what has been measured already is kept, and the price stays the same.
- The terms of service now say that the scores, analysis, recommendations and reports made from your questions belong to you, and can be used in any way, including in work for your clients. Before, the terms said they were ours and that you could use them.
- The terms now say that brandflare doesn't train AI models and doesn't let your data be used to train anyone else's, and that nothing from your workspace is published, sold or shared. The line that let us publish research from aggregated results is gone.
- The terms have a confidentiality section. Each side keeps the other's confidential information confidential, and for you that covers everything in your workspace.
- A change to the terms that takes something away from you is still emailed 30 days before it takes effect. A change that only adds to your rights, like these, takes effect when it is published.
- The privacy policy now lists every provider that receives your questions, and how: OpenAI, Google, Perplexity and Anthropic through their developer services, DeepSeek for the daily visibility check, and Bright Data, which asks your questions in the ChatGPT, Gemini, Google and Copilot apps. Facts, documents and answers go only to Anthropic and OpenAI, whose terms forbid training on them.
- Those calls to Anthropic and OpenAI are now made under zero data retention: the AI Gateway sends them only to providers that have agreed not to keep the request or its result once the call is done. It changes nothing you see, and nothing you pay.
- The owner of a workspace can now delete it, from Settings. Type the workspace's name to confirm, and it closes at once: checks and emails stop, nobody on the team can sign in to it, and any subscription stops at the end of the period you've paid for. Seven days later everything in it is erased for good: brands, questions, answers, history, the knowledge base and its documents, reports and the team. Until then the owner can sign in and undo it. The privacy policy says exactly what is kept after that, and why.
- Two new pages for security and procurement teams: Security, a one-page summary of how customer data is kept, reached and deleted, and Subprocessors, every provider that handles customer data, what it receives and where it is based. Both are linked beside the privacy policy and terms.
See the people AI sends to your website, from your own Google Analytics
- AI Traffic has a new tab, Visitors. Connect the brand's Google Analytics and it shows how many people reached your website from a link in an AI assistant's answer, which assistant sent them, and the pages they arrive on.
- Connecting takes a Google sign-in and, if your login can see more than one website, choosing the right one. brandflare asks only to read your Analytics: it can't change anything there and reads nothing else in your Google account. An owner or admin of the workspace connects it, from the Visitors tab.
- The last 90 days appear as soon as it is connected, and the page updates every morning. It shows visits from AI in the last 30 days and how that compares with the 30 days before, their share of all visits to your website, how many stayed, a line per assistant by day, and each assistant's visits beside how often it names you.
- Visits from ChatGPT, Perplexity, Gemini, Copilot, Claude, DeepSeek, Grok, Meta AI, Mistral, You.com and Poe are counted, along with anything Google Analytics files under its AI Assistant channel. Google's AI Overviews and AI Mode can't be told apart from an ordinary Google search, so they aren't included.
- AI Traffic is included on Plus, Pro and Agency, and during a Pro trial. The Crawlers tab shows what AI Crawlers showed before, and works with or without Visitors connected.
Questions in other languages, DeepSeek added to your daily visibility, restoring an archived brand needs a free place on your plan, and a report's branding can be changed and the report rebuilt
- You can now track questions in other languages. Add a question in Arabic, or any language your customers use, and it is asked in that language on every engine and answered in it. There is no setting: a question is asked in the language it is written in, and a brand can hold questions in more than one.
- Everything brandflare writes about those answers stays in English. Each question shows what it means in English underneath, and its topic, the brands named, how the answer describes you and any claim to review are in English, with the answer's own words kept beside a claim.
- An answer in another language has a Translate to English button. Press it once and the answer can be read in either language from then on.
- Your brand is counted however an answer spells it. An Arabic answer may write your name in Arabic letters and in several ways; each one counts as your brand and is highlighted in the answer.
- Whether a question in another language names your brand is worked out when you add it, so branded and unbranded questions are told apart in any language.
- Through the API, each question now carries in_english: what a question asked in another language means, or null for one asked in English.
- Export Report and one-off reports now handle other languages. The report is always in English: a question asked in another language is printed as it was asked with its meaning in English under it, a passage quoted from an answer has its translation under it, and the PowerPoint carries the English. A one-off report can have its questions written in another language too, chosen beside the market.
- Adding a brand and running a one-off report are laid out in numbered sections, and the market is now chosen in the brief, under Localisation, beside the question language.
- When you add a brand, the brief now asks for a question language. It starts on English, and you can choose Arabic, Chinese, Dutch, French, German, Italian, Japanese, Portuguese, Spanish or Swedish to have your first questions written in that language instead. It is only for the questions brandflare writes: you can add your own in any language afterwards, and mix languages on one brand.
- A one-off report's branding is now yours to set. The review step, before you run the report, shows the logo, banner and colour it will be drawn in, and you can upload your own logo or banner or choose the colour there. Until now a one-off report used whatever was found for the brand, and it couldn't be changed.
- A report can be rebuilt in new branding. After changing a brand's logo, banner or colour in Brand Settings, on the Branding tab, select Rebuild report: the PDF and the PowerPoint of its latest report are laid out again in about half a minute. Nothing is checked again, the findings and figures stay exactly as they are, and on a one-off report it isn't counted or charged as another report. Before, a report kept the branding it was first built with.
- Brand Settings on a one-off report now says that its branding can be changed, where it said the brand was archived.
- DeepSeek now counts towards your visibility. Each morning the questions that don't name your brand are put to it, and each answer is read for who it names and in what order. It has a line on your trends, a row under By engine and a column on Questions & Prompts.
- DeepSeek is measured for visibility alone, like Copilot and AI Overviews: no tone, accuracy or sources. It answers from what its model has learned, without searching the web. A brand's extra markets are not asked on it.
- Expect your Visibility and Flare score to move a little on DeepSeek's first day, because they now include an eighth engine. The change beside Visibility and Average position is measured on the engines that were measured both times, so a new engine doesn't read as a rise or a fall.
- Named by on Questions & Prompts and All engines on the overview count DeepSeek too, so both are out of eight engines on the questions that don't name your brand. Expect All engines to be lower than before.
- An archived brand keeps its place on your plan for 30 days, and restoring it in that time works as before. After the 30 days its place has gone back to your plan, so restoring it now needs a free one, the same as adding a brand. Before, a brand archived more than 30 days ago could be restored onto a plan that was already full. If there is no room, Restore says why.
- How it works now says what it costs. The free trial and the price of Pro are under the opening, what Pro includes is set out after the five steps, and the questions people ask most are answered at the end.
- AI Crawlers has moved. In a brand's sidebar it is now under AI Traffic, on a tab called Crawlers, and shows what it showed before. Links to the old page still work.
Your daily answers question by question, with Copilot and AI Overviews in the grid and competitors counted daily; visibility measured every day, on Copilot and Google's AI Overviews too, ChatGPT, Gemini and Google asked the way people ask them, with sources, and your brands in your own AI assistant, and unlimited seats and included one-off reports for agencies
- Agency now has unlimited seats, where it had ten. Everyone who works on a client can have their own sign-in, and nobody is counted or charged for.
- Agency now includes two one-off reports every calendar month. The review step and Settings › Reports show how many are left, and the Run button says whether a report is included or $99 before you press it. After the two, each report is $99 as before, and a report that can't be produced isn't counted.
- There is a new page for agencies, at brandflare.io/agencies: how a one-off report works as the audit you take into a pitch, what a client's report looks like, and what Agency costs.
- On a tablet, the site's header keeps the logo at full size and shows three links: Explore brands, Pricing and Agencies. The others are in the footer.
- Questions & Prompts now shows your daily answers. For every question that doesn't name your brand, each engine's cell is the latest answer, today's once the morning's check is in, and under it is a row of dots for the last 7 days: a filled green dot for a day you were named, an amber ring for a day you weren't. A question you hold shows seven green dots; one on the edge shows a mix.
- Copilot and AI Overviews have their own columns in the grid, on the questions that don't name your brand. A cell reads No overview on a day Google wrote none for that question. Named by counts them, so it is now out of seven engines on those questions.
- Opening a cell opens that day's answer, with the brands it named in order and the websites it used. A daily answer is read for who it names, so how each brand is described still comes with the weekly answers.
- Competitors counts the daily answers too, on all seven engines: who AI names, share over time and who's named first now move from day to day, and are drawn from several times as many answers. Share of voice on your overview and dashboard follows. Expect it to shift over the next few weeks as the daily answers fill its 30 days: every answer counts once, so the engines asked daily now weigh more in it than Claude, which is asked weekly. How AI describes each competitor still comes from the weekly check. For the first few weeks the head-to-head table leaves out whether a competitor is named more or less than before, until both spans it compares have daily answers.
- The change beside Visibility on a brand's overview now says whether it is real. AI can name different brands when asked the same question twice, so the figure moves a little when nothing has changed. From your last 7 days of answers brandflare works out how far chance alone moves it: a change bigger than that is green or red as before, and one no bigger is shown in grey. Each engine's change under By engine is coloured the same way, and What to do next no longer reports an engine naming you more or less often when the move is within chance. It starts once a brand has three days of daily answers.
- The change beside Visibility and Average position is now measured on the engines that were measured both times, 14 days apart. A newly added engine, such as Copilot or AI Overviews, is in today's figure straight away but no longer makes the change read as a rise or a fall on its own.
- All engines on the overview counts Copilot and AI Overviews as well, and reads each engine's latest answer. Expect it to be lower than before, because a question now has to be named by seven engines to count.
- A brand viewed across all its markets, and a brand's extra markets, show the weekly figures as before.
- You can now ask an AI assistant about your brands. brandflare has an MCP server, which is how Claude, ChatGPT, Cursor and other assistants connect to another product: add the connector address from Settings › API & MCP, sign in to brandflare when the assistant asks, and put your questions in plain English. How did the score move, which questions are we missing from, what is AI getting wrong about us, what should we fix first. It is on Pro and Agency, reads and changes nothing, and anyone on your team can connect their own assistant. The new guide to the MCP server has the steps for each assistant.
- The API is now part of Pro as well as Agency, and of a Pro trial. Make a key in Settings › API & MCP to read your brands from your own dashboards and spreadsheets.
- The API returns four more things for each brand: its questions with what every engine's latest answer did with the brand, the accuracy claims with why each is wrong and what is true, the recommendations with their steps, and the sources AI cites. An assistant connected through the MCP server can read the same.
- The pricing page is up to date with all of this: Pro and Agency list the API and AI assistant connector, Plus, Pro and Agency list Knowledge from your site and files, and the answer on the free trial says what starts when it converts. Agency's description now names what sets it apart: unlimited seats and reports in each client's colours.
- Your visibility is now measured every day. Between your weekly checks, the questions that don't name your brand are put to ChatGPT, Gemini, Google Search and Perplexity each morning, and each answer is read for who it names and in what order. It starts on 5 October for every brand you track, on every plan and during a trial.
- On a brand's overview, Visibility and Average position, their trend lines and the Flare score now move from day to day, and so do the cards on your dashboard. The Updated time at the top of a brand includes the daily measurement, and the date beside it is your next full check.
- Tone, accuracy, sources, Brand Narrative, alerts and the weekly summary still come from the weekly check, which reads every answer in full. Claude is still asked weekly, as are questions that name your brand and a brand's extra markets, and a brand viewed across all its markets shows the weekly figures.
- Through the API, a brand's visibility, average position, Flare score and share of voice can now change daily. The API guide says which figures are daily and which weekly.
- Two more engines count towards your visibility from 5 October: Microsoft Copilot, and Google's AI Overviews, the answer box at the top of an ordinary Google search. Each morning the questions that don't name your brand are put to both, from your brand's market, and each answer is read for who it names and in what order.
- On a brand's overview each has its own row under By engine and its own line on the trends, and both count in Visibility, Average position, share of voice and the Flare score. Expect those figures to move on the first morning, because two more engines are in the average.
- Copilot and AI Overviews are measured for visibility alone. Tone, accuracy and sources still come from the weekly check of ChatGPT, Gemini, Google Search, Perplexity and Claude, and a brand's extra markets are not asked on the two. Google doesn't write an AI Overview for every question, so your visibility there is the share of the overviews it did write that named you.
- brandflare now asks ChatGPT and Gemini the way a person does: in the app itself, with web search on, from your brand's market. Until now both were asked without web search, so their answers came from what the model had learned and carried no sources. This starts with each brand's next check.
- Google Search is now Google's own AI Mode, the answer Google itself gives in search. Until now it was Gemini with Google's search switched on, which stood in for it.
- All three now come with the pages they cite, so Sources shows what ChatGPT, Gemini and Google read in your category, alongside Perplexity.
- Expect your figures on these three engines to move on the first check after this change, because these are different answers: they follow what the web says today. It also means a change to your own pages can show up in their answers within weeks, where before it had to wait for a new model.
- A new brand's first check takes a few minutes longer, because the apps take longer to answer than the models alone.
- Claude is still asked as a model, without web search. The guide on how checks work says how each engine is asked.
- Emails about what is new on brandflare now carry an unsubscribe link, and Profile › Alerts has a Product updates switch for them. Each person chooses for themselves.
- Every unsubscribe page now also offers to stop every email from brandflare to your address at once: alerts, weekly summaries, getting-started emails and product updates. Profile › Alerts shows when that is so, with a button to start them again. Emails about a payment, your sign-in or something you asked for are still sent.
- We now use Google Analytics to count visits to the public site. It records pages on the public site and never pages inside an account. The Cookies and Privacy pages list it.
Brand knowledge, a pricing page that covers the newest features, and public pages that look the same signed in
- Brand Settings has a new Knowledge tab, in beta, in place of Facts. Paste a link, drop in a PDF, Word or text file, or write what is true, and brandflare reads it into facts your answers are checked against. Each fact shows the source it came from and the passage that says it, word for word.
- On Plus, Pro and Agency, brandflare reads your brand's own website without being asked, so the Knowledge tab has your own facts in it before you add anything. Brands you already track have theirs read over the coming days.
- There is no longer a list of facts to verify. A fact counts as soon as it is read, and you remove any that are wrong. The one thing you are asked to settle is a disagreement: when two of your sources say different things, both are shown side by side and you keep the right one.
- A claim in Answer Accuracy that contradicts one of these facts now shows the proof: the passage from your own page or document, with a link to it, beside what AI said.
- Answers are checked against up to 300 facts for a brand, up from about twenty. Expect more claims to review after your next check, most of them details such as prices that your own pages state and research had missed.
- The Onboarding checklist follows: its fact sheet step is now Check what we know about you, and on Plus, Pro and Agency there is a new step, Add a page or document of your own.
- When you're signed in, the public pages now open as the public site shows them, without the app's sidebar round them: Pricing, How it works, Get in touch, the guides, the blog, the changelog and the legal pages. A Dashboard button at the top, where visitors see Sign in, takes you back to your account. Explore and brand profiles still open inside the app.
- The questions on the pricing page now cover the API and one-off reports, both part of Agency: what the API returns and where it can be read from, and how a report on a brand you don't track is made and what it costs.
- The pricing page now says that extra brands can be added on any paid plan, one at a time for $69 a month each, from Plans & billing. It used to mention Agency only.
- The answer on questions that name your brand is corrected. They are left out of your visibility, but they are asked on purpose, to check what AI says about you when someone asks directly, and the overview compares the two.
- Agency's description on the pricing page and at checkout now names what sets it apart: client-ready reports, the API and ten seats.
- In the knowledge base, the plans page lists extra seats and extra markets, and the plans and billing pages say that cancelling is something you do yourself, and that it leaves you on Free with everything you've measured.
One-off reports, branded vs unbranded, an API for agencies, two-factor authentication, and sign in with Google
- Agency workspaces can now buy a one-off report on any brand for $99, without tracking it. Choose Run one-off report in the menu under your avatar: we research it, you read and change the competitors and questions it will be asked, and only then run it. The brand is checked once across all five engines, and you get its results in the app with the PDF and PowerPoint. It doesn't use a brand slot, and Start tracking turns it into a tracked brand whenever you like.
- Settings has a new Reports tab that keeps every report your workspace has made, newest first, so earlier ones are easy to find and download again.
- A brand's overview has a new card, Branded vs unbranded. Unbranded questions don't name your brand, so they show whether AI thinks of you; branded questions name it, so they show what AI says about you when someone asks directly. The card sets the two side by side on your latest answers: how often you are named, named first and recommended, the tone, and the share of answers that carry a false claim. Both kinds of question were already part of every check, so nothing more is asked.
- The tabs on Questions & Prompts are now All, Branded and Unbranded, and each branded question carries a tag. They replace Not always named, Named everywhere and Paused: the list still puts your weakest questions first, and paused questions are at the end of it with their Resume button.
- The All engines figure on the overview now counts unbranded questions only. Nearly every engine names you when the question itself does, so those questions made it read higher than it should. Expect it to be lower than before; your Visibility and Flare score are unchanged, because they never counted branded questions.
- The brand report has a new page after the score, Branded vs unbranded, with the same comparison, a written reading of it, and both kinds of question engine by engine. The PowerPoint edition has a matching slide.
- Agency workspaces can now read their brands' figures from their own dashboards and spreadsheets. The API returns each brand's Flare score and five factors, its score history, how each AI engine treats it, and share of voice against competitors. It reads and changes nothing.
- Keys are made in Settings, on a new API tab. A key can read all your brands or one brand only, so a key in one client's dashboard never shows another client's figures. The same tab shows the real answer for your brands, with the request written out for you.
- Every address can also return a table, so the figures drop straight into Google Sheets with one formula, and from there into Looker Studio. The API guide in the knowledge base covers the first request, Sheets, every address and every error.
- You can now add two-factor authentication to your sign-in, under Profile, Security. Scan a code with an authenticator app such as Google Authenticator, Microsoft Authenticator or 1Password, and from then on signing in asks for a six-digit code from the app as well as your password. You get backup codes for a lost phone, and can move to a new phone or turn it off from the same page.
- Owners and admins can require two-factor authentication of everyone on a workspace, from Settings, Team. People already on the team get an email and 7 days to set it up, with a reminder at 3 days; after that the workspace is closed to them until they do, and nothing is removed. Anyone who joins later sets it up the first time they sign in. The Team page shows who has it, and an admin can reset it for a colleague who has lost their phone.
- You can now create an account and sign in with Google: one click on Continue with Google, no password to choose and no code to wait for. If you already have an account under the same email address, Google signs you in to that account rather than making a second one. Email and password still work as before, and Profile lists Google under Connected accounts.
- A brand whose check is late no longer shows an "Overdue, we're on it" label in its header. The dot beside Updated still turns amber, and Next update still gives the date of the next check.
Clearer billing dates on Subscription
- When your plan is free for a while, Subscription now says so and gives the date of the first renewal charged at the full price. It used to call this a free trial and show the day the discount ended, which could be a month early.
- Any brand slots given to you free are now counted with your extra slots and marked as free, so the page shows exactly what is on your bill.
The company behind brandflare in the footer
- The footer on every page now names the company that operates brandflare: NONE OF THE ABOVE L.L.C-FZ.
The Meta Pixel on the public site
- We now also use the LinkedIn Insight Tag, to measure our advertising on LinkedIn. Like the Meta Pixel it records pages on the public site and never pages inside an account. The Cookies and Privacy pages list it.
- When someone creates an account or starts a subscription, our server now tells Meta as well as the browser, so the count is right even where an ad blocker stops the pixel. The email address is hashed before it is sent. If a visit came from one of our campaigns, a new account remembers which one. The Cookies and Privacy pages say so.
- We now use the Meta Pixel to measure our advertising on Facebook and Instagram. It records pages viewed on the public site and when someone starts signing up, creates an account, adds a brand or goes to checkout, but never pages inside an account. The Cookies and Privacy pages list it.
Answer Accuracy by topic, and the fact sheet in Brand Settings
- The brand picker in the sidebar now shows just your brands, without a red count of claims to review beside them.
- Claims to review now opens on topics: the subjects AI gets wrong about your brand, such as pricing, products or leadership. Each topic shows how many claims are waiting and how serious they are, which engines say them, how many facts on your sheet cover it, and whether it had more or fewer claims at the last check than at the one before. Open a topic to review its claims one at a time, or review everything in one queue as before.
- Topics are set for you. When AI starts getting a new subject wrong, it gets a topic of its own, marked New until someone opens it. A topic your fact sheet says nothing about is marked, with a link to add the facts.
- The fact sheet has moved to Brand Settings, as a fourth tab called Facts, with its facts grouped under the same topics. A link beside the severity filters on Answer Accuracy opens it, and old links to it go to the new place.
- Claims to review now holds only what AI is still saying. A claim that none of the last three checks heard leaves the queue by itself and moves to Dismissed as No longer said. If an engine says it again, it comes straight back.
- The brand report has a new page after Accuracy: the false claims by topic, each with a line on what the engines get wrong, how many claims and how many are serious, the worst example, and how many facts on the fact sheet cover it.
Faster pages
- Brand pages and the dashboard load faster: they now ask for their figures together instead of one after another, and no longer ask for the same figure twice.
- Pages download less. Brand logos arrive at the size they are shown instead of as large originals, which makes the home page and Explore much lighter, and pages that draw no charts no longer download the charting code.
- The Find your brand box on the home page fetches the list of brands when you first click into it, rather than with every visit to the page.
- The home page and Explore now open almost instantly: the index figures are worked out every few minutes rather than for each visit. Answer Accuracy and Competitors are quicker for brands with a long history, and rivals with identical figures are listed in name order.
- We now measure how quickly pages load for visitors with Vercel Speed Insights. Like our analytics it is cookie-free, and the Cookies page says so.
Extra markets: one brand, checked from up to five countries
- On Pro and Agency, a brand can now be checked from up to four more countries as well as its home market. Each extra market asks the same questions, up to 40 of them, as someone located in that country, on the same engines and the same weekly schedule. An extra market is $49 a month, or $490 a year, per brand.
- Add one under Brand Settings › Markets: choose Add a market, pick the country and pay without leaving the page. You pay only for the rest of the current billing period, the new market's first check starts straight away, and from then on it is checked on the same day as the brand. A market you stop keeps running until the end of the period you paid for, and can be kept before then.
- Brand Settings now has three tabs: Branding, Markets and Competitors. The brand's home market and its competitors each show as a list with a pencil to change them, and the brief a brand was added with is under Competitors.
- A brand with more than one market has a market picker beside its name. All markets, the default, combines them; choose a country to see every page, the Flare score in the header and the numbers in the menu for that country alone. Its flag shows beside the brand while you do.
- The new Markets page puts a brand's markets side by side: the Flare score, how often it is named and in what position in each country, which engines name it where, the questions where the countries' answers differ most, and who else is named in each, with the brand and its nearest rival highlighted.
- Recommendations that apply to only some markets show those markets' flags.
- The brand report covers every market. After the home market's pages come a market comparison and a page for each market, with its figures, its biggest gaps and a short note on what sets it apart. The Agency PowerPoint has the same, with each market's flag.
- The billing page lists each brand's extra markets and what they add to your next charge, and the pricing page shows extra markets beside the plans. Workspace menu items are a little smaller, so every section fits on screen without scrolling.
- On Pro and Agency, the getting-started emails now include one on markets, in the second week: how to add one, or, once you have, how to read the comparison.
A brief when you add a brand
- Adding a brand has a new, optional step: a brief. Name the competitors it must be compared with, and write a few lines on what the research should focus on, such as who is asking, the territories to explore and anything to leave out. The research, the competitors and the questions follow it.
- The competitors you name are always included, first, each with its website so the research looks at the right company. Research adds others only if your plan has room for more.
- A brand from the index that is given a brief is researched for your workspace instead of copied, so its questions follow your brief rather than the index's. Without a brief it is added as before.
- The brief stays with the brand, under Brand Settings › Tracking, so you can always see what its questions were built from.
- Every way of adding a brand now goes through the same steps: Add brand in the menu, your first brand, Add to your workspace on a brand's public page, and the brand you looked up before signing up, which now opens at the brief with the brand already chosen. Nothing is added until you start the research.
Brands added from the index are checked on all five engines
- A brand you add from the index now gets Google Search answers straight away, alongside the other four engines. Until today its first results came only from ChatGPT, Gemini, Perplexity and Claude, because the index is checked on those four, and Google Search waited for the brand's first weekly check.
- You now hear the moment a brand added from the index is ready, by email and in the bell, as you already did for a brand set up from scratch.
How it works
- There is a new How it works page, in the top menu where Learn used to be: five short steps from looking your brand up for free to fixing what AI gets wrong about it. The guides are still in the footer, and linked from the end of the page.
- It spells out the difference between the public index and your own brands: brands you add are private to your workspace and are never listed on Explore.
- Plans & pricing is now just Pricing, in the top menu and the footer.
Get in touch
- There is now a Get in touch page, in the top menu and the footer, for questions about plans, agency partnerships, press or getting into your account. You don't need to be signed in, and we reply by email.
- Pages that used to ask you to email us, including the suspended and closed account pages and Enterprise on pricing, now open the form with the right reason chosen.
The index, checked monthly
- Brands in the index are now checked once a month, each on its own day, and their Flare score is built from the last month of answers. The score, the change beside it, Movers on Explore and the brands you follow all measure against a month earlier. Brands you track are still checked every week.
- More brands in the index open on a real header image: where the usual source had none, we now use the brand's own banner, found elsewhere, instead of generated artwork. It stays with the brand when you add it to your workspace.
- Brands in the index are asked on four engines: ChatGPT, Gemini, Perplexity and Claude. Google Search is kept for the brands you track, where it is checked every week alongside the other four.
Following, on the home page
- The home page now walks through five steps rather than four: Ask, Measure, Judge, Recommend and Follow, the last showing what a watchlist of brands you don't track actually looks like.
- The AI Flare Index on the home page is the top seven brands in order, instead of seven picked down the length of the ranking. It no longer says when the index was last checked, since "live" already covers it.
- The strip of brand logos under the home page's opening now shows sixty brands from across the index, the best in each category first, instead of every brand in it. The page loads far faster: it was fetching over two thousand logos to show the ten on screen.
Leaving a plan, and seats on Pro
- Cancelling is now something you do yourself, on Plans & billing, instead of writing to us and waiting. Your plan runs to the end of what you have paid for, and on a trial that means the day the card would have been charged, so nothing is taken.
- On the way out we offer you Plus first, because most people leaving Pro are leaving the price rather than the product. Taking it moves you down at the next renewal with nothing charged today.
- Cancelling now leaves you on Free rather than locked out: every answer, score and claim we have gathered for your brands is still there, you can still follow brands and read the index, and coming back picks up exactly where you left off.
- You can say why you are leaving on the way out, and we read it.
- Pro covers three people, and you can now buy more seats, one at a time, for $69 a month each. The seat is yours the moment you pay for it, so the invitation you were about to send goes out straight away.
- Seats are bought and given back on Plans & billing, under the seat meter, the same way as extra brands. A seat you give back is paid for to the end of the period, so the person in it carries on until then and your next invoice simply charges less.
- When every seat is taken, the Team page now offers you a seat rather than only a bigger plan.
- A month after a trial ends without a plan, we write once with what has happened to the brand since — its last Flare score and anything AI is still getting wrong about it — and half off the first three months of Plus or Pro. It is the only note after a trial ends, and there is nothing after it.
Follow brands, and hear the moment your first results land
- A brand you track always opens as your own. Before, the index version of it was still reachable, showing a teaser of what you already pay for and offering to let you follow it; now it takes you to your brand instead.
- New accounts hear from us once, the day after signing up, about following brands and what tracking your own would show instead.
- "Follow a rival you don’t track" joins the onboarding checklist, since following lives on the index and nobody tracking a brand would otherwise come across it.
- You can now follow brands from the index without tracking them. Each one appears on your new Following page with its Flare score, how that score has moved in a fortnight, and its category, and you get an email every Monday on what changed. On the Free plan you can follow three brands; on a paid plan, as many as you like, which is a way to keep an eye on a rival you don't pay to track.
- The Monday email leads with the biggest mover, lists where each brand you follow stands, and shows one real thing AI is getting wrong about one of them. One click at the bottom stops it, without signing in.
- Big moves on a brand you follow reach the bell in the header too: a new false claim, a rival passing it, an engine dropping it, or the score moving.
- When a brand's first check finishes, we now tell you straight away, by email and in the bell, with its Flare score, how often AI names it and the first thing it gets wrong. Setting a brand up takes about half an hour, and until now nothing marked the end of it.
- Following a brand costs no checks: every brand in the index is measured anyway, so watching one changes nothing about your plan's allowance.
Answer Accuracy: five ways to deal with a claim
- Every brand section now leads with its tabs rather than a heading block: the section's name, an icon and, at the end of the row, a (?) that says what the page and the view you are on are for. Sections with a single view, like Competitors and Crawlers, read the same way. It puts the work itself at the top of the page instead of below a paragraph.
- A claim used to leave you two buttons, and both made the card vanish with nothing to show for it. There are now five, and each says what it does. Watch: AI has this wrong, follow it until the engines stop saying it. Fix it: the same, and it becomes a recommendation carrying the claim, what's true, who says it and the sites those answers cited. Dismiss: you've seen it and don't want it in the queue. Add to fact sheet: the claim is true after all, so it becomes a verified fact and stops counting against you. Not about us: the flag should never have been raised.
- Two new tabs. Being watched follows the claims you decided are real, marking each as still said or stopped, with when it was last heard — so you can see whether a fix worked. Dismissed holds the rest, says what each decision did to your score, and can put any of them back in the queue.
- Every decision can be undone, and a claim you have decided no longer comes back into the queue at the next check. Dismissing one says plainly that it still counts against your accuracy while AI keeps saying it.
- The claim card now says how widespread a claim is ("in 3 of 87 answers") and when it was first said, and Read the answer sits beside the explanation.
- The Answer Accuracy pages lead with their tabs, each with an icon, and a (?) at the end of the row explains the page and whichever tab you are on. Your fact sheet is named after the brand.
- The onboarding checklist now ticks when you review a claim, and opening your fact sheet counts as checking it.
A bell for what changed
- On Sources, the outreach lists now lead somewhere: View pages on any site opens the pages of it AI actually read, and each page opens the questions that led there, with the answers. Coming back returns you to the outreach list. A website and a number was never something you could act on; a page and a question is.
- Every engine logo now says which engine it is when you hover over it, or reach it with the keyboard.
- There is now a bell in the header, with a count of what you haven't read. It holds what each check found about your brands: a rival passing you, an engine that stopped naming you, a new false claim, a negative story taking hold, sites AI has started or stopped citing. Each one names the brand, says what happened and when, and opens the page that explains it.
- It carries everything, whatever your alert emails are set to, because a note in the app costs you nothing: brands set to the weekly summary, or to nothing at all, still show their changes here. What you see matches the emails in one way — a change only appears when it is bigger than the wobble between checks.
- Our replies to your support requests appear in the bell too, in place of the dot that used to sit on your avatar, along with changes to your account: a trial ending, a payment that didn't go through, a plan change, and colleagues invited or joining. Account notices go to the people who run the account.
- What you have read is yours: clearing your bell leaves your colleagues' alone. See everything opens the full list, where Mark all as read clears the count.
- The bell starts with the last month of your checks already in it, marked as read, so it is not empty on the first day.
One free trial each, tidier emails, and Brand Settings
- The first time you open a brand after its first check, a short tour shows you around: the Flare score and the five figures under it, then each section in the menu, in under a minute. Skip it at any point, and replay it from Take the tour at the top of any brand page, or hover over that and dismiss it. Anyone invited to your workspace later gets it too. It runs on a computer, not on a phone.
- The tour ends on an Onboarding checklist in the corner of the screen: the first things worth doing for the brand, from reviewing its questions and confirming its competitors to exporting a report. Each ticks itself off once it's done, and Dismiss puts it away.
- Export Report now sits with Brand Settings at the bottom of a brand's menu, below the sections that show results.
- If you stay on a brand while its first check runs, the page now fills in by itself when the results arrive, without a reload.
- Every brand has a new Brand Settings section, at the bottom of its menu. Choose the brand's colour there if the one we read from its site is wrong, or if we couldn't find one: it is used for the brand's header, its reports and the emails about it. The brand's location, logo and banner can be changed there too.
- You can now change who a brand is compared with. Brand Settings now has two tabs, Branding and Tracking, and shows only the settings, without the brand header. Tracking holds the brand's location and its competitors, and Edit competitors on the Competitors page goes straight to it: add a rival with its website, remove one, or correct a name, up to your plan's limit. Past share of voice and scores are recalculated with the new list, and a newly added competitor is fully counted from the next check. Until then, the Competitors page names it and says when that is.
- Setting up a new brand now finds as many competitors as your plan compares, rather than stopping at four, and the last step shows who they are with a link to change them.
- Uploading a logo for a brand we hold no colour for now takes the brand's colour from the logo, so its header and reports are no longer drawn in grey.
- Every email from brandflare now shares one look: the team invitation, and the notes our team sends when we change something on your account (a plan change, a suspension, a closure, a password reset, a new member), are drawn the same way as your alerts and reports, with the logo inside the message. Every one of them can be replied to, and replies reach our support team.
- The Pro free trial is now once per person, once per card, and a few per company. If colleagues at your company have already had theirs, the pricing page says so and offers Pro from today, or you can ask one of them to invite you to their workspace. A card that has already started a trial starts a paid plan instead, charged from that day. Companies that are really a group of separate businesses can ask support to have this lifted.
- Sign-up no longer accepts throwaway email addresses, or the same address with a plus tag added. Existing accounts are not affected.
- When your brand slots are full, Add a brand no longer counts recently archived brands as ones you are tracking. It says your slots are in use and when an archived brand's slot frees up.
- Signing out is now one clean step: a brief "Signing you out" screen, then the homepage, instead of the app emptying out around you first.
- We fixed a mistake in how visibility was counted, and it read too high. An answer that named your brand alongside several competitors was counted once for each brand it named, instead of once. Visibility now means exactly what it says: the share of answers that name you. Most brands read a few points lower, and more in busy categories where answers list many rivals. The drop is the correction, not a change in how AI talks about you, and past figures have been recalculated the same way so trends still compare like for like. Share of voice was already counted correctly and is unchanged.
- Accuracy now weighs false claims against every answer about your brand, including answers to questions that name you, since that is where many false claims appear. Before, those answers were left out of the count, which made some brands look less accurate than they are. The Flare score moves slightly with both fixes.
- Alerts now only go out when a change is bigger than chance. Each check asks every question once per engine, so a brand's figures wobble a little from one check to the next, most of all on a small set of questions. An alert about visibility, share of voice, a rival passing you, an engine dropping or picking you up, or sentiment now waits until the move is clearly larger than that wobble, given how many answers stand behind it. A Flare score alert needs one of those to have really moved, or a new false claim. Expect fewer alerts, and ones worth reading.
- A negative theme in Brand Narrative is only reported as new the first time it appears. Before, every rebuild of the narrative could report the same themes again.
A report you can send on, and support inside brandflare
- Your first two weeks now come with a short series of getting-started emails, whether you begin with the free trial or go straight onto a plan. Each is about one part of brandflare: reading your first results, what AI gets wrong about you, choosing your questions, Brand Narrative and Sources, Export Report, and Recommendations. They quote your own figures, show a picture of the section they are about, and only cover what your plan includes.
- You are not told what you already know: an email is skipped once you have done the thing it is about, you never get two in a day, and they go to the owner of the account only. One click in the footer stops them, and a switch under Profile, Alerts turns them back on, carrying on from that day rather than sending the ones you missed. Emails about your trial or a payment are separate, and carry on as before.
- Anyone who goes straight onto a paid plan now gets a welcome email the moment they subscribe, with where to start.
- Agency now includes Premium Support: requests from an Agency workspace are answered first. There is nothing to switch on. Raise a request from Customer support as usual and your plan travels with it.
- The Pro free trial is now 14 days rather than 30, for trials started from today. Everything else about it is the same: the full plan on one brand, a card taken at the start, nothing charged until the trial ends, and an email a week before and again three days before the first payment. A trial already under way keeps the end date it started with.
- Every brand has a new Export Report section: a full audit of how AI answers questions about it, written up and laid out as a PDF. A cover with the single most important finding, a one-page summary that stands on its own, the Flare score and its five pillars, the questions the brand is missing from with a play for each, every question engine by engine, share of voice against the rivals you track, what the engines actually said, how they describe the brand, the false claims that matter most, and the sites the answers cite.
- Press Generate and it takes about a minute. You can watch it being built or leave: it downloads on its own if you stay, and we email you a link either way. The link asks you to sign in, because a report is only for people on your workspace.
- A report is of a check. Once a brand has been checked again, you can build a new one, and it opens with what changed since the last: the score, how often the brand is named and named first, false claims, and the questions it gained and lost. Until then the button downloads the report you already have, because there is nothing new to say.
- The findings are written for you from that check's figures and checked against them. Every number, table and chart is computed directly from the judged answers, and the last page says how.
- Included on Plus, Pro and during a Pro trial. On Agency the report is drawn in the brand's own colours, logo and header image, ready to hand to a client. Starter shows what it contains and how to get it.
- On Agency the same report also downloads as a PowerPoint: a slide to a section, with charts and tables you can edit and restyle and the detail behind each slide in its speaker notes, for lifting into your own presentations. It is built from the same findings, so there is nothing to choose up front: generate once, then pick PDF or PowerPoint when you download.
- Customer support in the avatar menu now opens a Support page instead of your email. Send us a question, a billing query or a problem from a short form, and the whole conversation stays on the request's page where you can find it again.
- We reply by email and on the page. The email carries our full answer and a link back; you answer on the page, not by replying to the email.
- When we've replied, a dot on your avatar and a count beside Customer support in its menu say so, on every page.
- The Support page lists your open requests in a table, each with its number, topic, when it was last updated and whose turn it is: awaiting support, or awaiting your reply. Mark one as resolved yourself when it is dealt with. Once we've replied, a request closes after seven days without an answer, with an email two days before. Replying to a closed request reopens it.
- Owners and admins see every request from their workspace, so a billing question a colleague asked isn't asked twice. Editors and viewers see their own.
- Your plan and the page you were on are sent with a request, and the brand when you were looking at one, so you don't have to explain where you were.
- On a phone, Add brand has moved into the menu, at the top, instead of taking a row of its own under every page title. Until you track your first brand it stays where it was, since it is the one thing to do.
- Asking for a downgrade or a cancellation, help connecting AI Crawlers, or more brands on a managed account now opens a request you can follow, already filled in, rather than an email.
- Our logo now travels inside every email we send, so it shows in Gmail and on a phone instead of as a broken image, and stays readable when your mail app switches to dark mode.
- The page you are on is now marked in the menus with the same coral-to-pink sweep as our buttons, in the sidebar and in the bar across the top of the public pages. In the sidebar it is no longer in bold.
- The emails that carry your sign-in code, a password reset code or a sign-in link now look like the rest of our mail: the new logo, our type and colours, and the code set large enough to read at a glance. They still say where and when the request was made.
- The brandflare icon in your browser tab now carries the same coral-to-pink sweep as our buttons, instead of flat pink.
- A brand we hold no header image or colours for now has a plain light grey band across the top of its page, in place of pink ruled lines that had nothing to do with the brand.
- Brands now have a market: the country their customers are in. AI answers differently depending on where the person asking is, and asked with no location it mostly answers as if for someone in the United States. Set a market and every check is asked as someone there, so a UK brand is measured on what UK customers are told.
- Choose the market when you add a brand. Its questions are then written for customers there, in their currency and with the shops and rivals they know. Global (no location) is still there for brands that sell everywhere.
- Brands you already track have a Set a market chip at the top of their page, beside the last update. It applies from the next check, and scores may step on that date because the answers come from a different place. Each answer now says where it was asked from.
- Settings has a Default market for the workspace, so every brand you add from now on starts there. Brands you already track keep their own.
- Sources has a new Your pages tab: the pages of your own site that AI cited in the last 30 days, how many times, by which engines, and what the questions were about. Open a page to see the questions that led AI to it, and click one to read the answer. If AI isn't citing your site at all, it says so.
- Any website under Most cited now opens to show which of its pages were cited, so you can see the exact article or thread AI is reading, not just the site.
- Every question now has a topic, a word or two such as Pricing or Delivery, shown on Questions & Prompts, where you can filter the list by it. A brand's questions are sorted into six topics at most, and the filter shows the busiest few with the rest under Other. New questions you add are given a topic a moment after you save; there is nothing to pick.
- Export Report lists the pages of your own site the engines cited, in the Sources section of the PDF and on a slide of its own in the PowerPoint.
- Signing in now goes from the form to a short Signing you in screen and then to your dashboard. It used to flicker on the way: the form vanished and left an empty card, then the dashboard showed for a moment under the public site's header before the app took over.
- The privacy policy and terms of service are complete, and there is a new cookie policy listing the few cookies brandflare sets. They name the company behind brandflare, who handles data for us and where it is stored, how the free trial and renewals work, and which law governs the agreement. All three are linked from the footer.
- Paying no longer sends you away to another site. Checkout is now a page inside brandflare: what you are buying and what it costs on the left, the card form on the right. Starting a trial, it shows $0 today, the date and amount of the first charge, and when we will email you before then. Your card details still go straight to Stripe and never reach us.
- The email you get when a trial starts now says what to do first. If you have not added a brand yet, that is step one, with a button straight to it and a subject line that says so. If you chose a brand on the way in, it tells you we are already setting it up and there is nothing to do.
- The product docs under Learn now cover the whole of brandflare: a page for every section of a brand, from Questions & Prompts to Export Report, plus markets, alerts, teams and roles, managing brands, the free trial and billing, and Support. They use the same words as the screens: questions, checks and claims to review.
- The pictures at the top of each blog post are redrawn in the brandflare pink with the current logo, so the blog matches the rest of the site. The Changelog link in the footer now sits under Product.
- The blog has a name: Quoted. It is at the same address, brandflare.io/blog, with the same posts and the same RSS feed. The monthly engine report sign-up box has come off its pages; nobody had been sent anything.
brandflare turns pink, with a new logo
- A new brandflare logo: a quote mark over the name. It is in the sidebar, on the home page and sign-in, and at the top of the emails we send.
- The accent colour across brandflare is now pink rather than orange: links, buttons, active tabs, highlights and the emails we send. Nothing has moved and nothing works differently.
- Score colours are unchanged. Amber and orange still mean a moderate or low flare score, and warnings are still amber.
- The home page shows the product as it is now: real brand pages you can flick through, and the actual Questions, Answer Accuracy and Recommendations screens in place of drawings. It also says plainly what brandflare does, and that brands are checked every week.
- Emails switch to a white logo on a dark page in mail apps that tell us they are in dark mode. Gmail does not, so there the emails stay light.
- On sign-in and sign-up, boxes you type into have a visible edge before you click into them, and a thinner pink one when you do.
- A brand you add from the index now arrives with its header image and brand colours, where some used to open on a blank band.
- Brand descriptions under a brand's name are one short sentence, and the Flare score stays beside the name instead of dropping underneath on narrower screens.
- On a brand's public profile, the button to add that brand now sits under its description. The header keeps the one Add brand button it has everywhere else.
- The pricing page leads with Pro and its free trial, with Free, Starter, Plus and Agency beneath. Each plan's list now says exactly what you get, with what it doesn't include at the end, and says plainly that Agency can add brands for $69 a month each.
- Click a brand without an account and the sign-up page now says which brand you were after, and takes you straight to its profile once you're in. Signing in instead of signing up goes to the same place.
- Explore: can't find your brand? The link now takes you straight to adding it yourself, with the name you searched for filled in, instead of a form asking us to add it. Popular categories are listed alphabetically.
New plans, and a lower way in
- Starter is now $29 a month: one brand, 15 questions, checked weekly across all five engines.
- Plus is new, at $79: the whole product on one brand — 40 questions, five competitors, Brand Narrative, AI Crawlers and alerts. The step up from Starter is depth, not a second brand.
- Pro stays at $199 for three brands, and Agency at $399 now covers five brands rather than ten. Anyone already tracking more keeps every brand they have.
- Need more brands than your plan covers? Add them one at a time for $69 a month each, from Plans & billing or from the add-a-brand screen when you reach the limit. Charged pro rata from the day you add one, and removable the same way.
- Yearly billing still gives two months free on every plan.
- Reach your brand limit and the add-a-brand screen now says so and takes the payment on the spot — the next plan up on Starter, Plus and Pro, an extra brand on Agency — with the exact amount due before you confirm, and the wizard opens straight afterwards.
- Brands are bought one slot at a time, from the usage meter on Subscription: Add a slot says the price, takes the payment and opens the add-a-brand wizard. Remove a slot sits next to it once you have one.
- Your plan now shows what the subscription actually comes to, the plan plus any slots, rather than the plan's own price. Where a discount or an agreed free period applies, it leads with what is actually charged next and says when full price starts.
- A slot bought during a trial is charged there and then. The trial is free time on the plan, not on brands added alongside it.
- A slot you have given back can't be filled with a new brand while it runs out — the brand would stop the day the slot did. Adding a brand reinstates it instead, at no cost, since it is already paid for.
- Reaching your brand limit no longer replaces the add-a-brand screen with a page that only sells. Where there is something to pay, it becomes the wizard's first step — Add brand quota, with the amount due today and what it renews at — and naming the brand follows it. Where there isn't, because a slot you had given back is coming home, a line at the top says so and nothing else changes.
- An extra brand is a month, not a number of days. Remove one and it keeps working to the end of the period you have paid for, then stops and is never charged again — nothing is refunded for the days left, and nothing extra is charged. Change your mind before that date and Keep them puts it back at no cost.
- An archived brand no longer keeps you paying. Its slot is still held for thirty days, so the room does not come back straight away, but you can hand the paid brand back the moment you stop tracking it.
Try Pro free for 30 days
- Pro now starts with a free trial: thirty days of the full plan, with a card taken at sign-up and nothing charged until the trial ends. Cancel any time before then and you pay nothing.
- A trial tracks one brand, with everything else Pro includes — 40 questions, 5 competitors, Brand Narrative, AI Crawlers, alerts and three seats. Converting to Pro raises it to three brands.
- Every page shows how many days are left, and Plan & billing says the date of the first charge.
- One trial per account, and it needs a confirmed email address. Accounts already on the free plan can start one too.
- Four emails across the month: a welcome with where to start, a note a week before the end with what has been found so far, a reminder three days out naming the exact amount and date of the first payment, and a note when the trial finishes either way. Nobody is charged without being told first.
- A plan changed during a trial keeps the same end date, so you can move to Starter before the first payment without losing days.
- Seen enough? Start your subscription early from Plan & billing, or from the add-a-brand screen, and the trial ends there and then: you're charged the first period and go from one brand to three.
- Brands can't be archived while a trial runs. Archiving frees the slot, and a trial has only one.
- While a trial is running, Plans & billing marks the plan it is on and says one brand while your trial runs, then three — rather than promising three next to a meter reading one.
- Plans and billing are now two tabs. Subscription holds your plan, what you are using of it, and changing plan; Billing holds your invoices and the card we charge, shown as a card.
- Adding your first brand opens with a welcome and how long your trial has left, above the list of what happens next. The page is titled Add your first brand while it is your first.
- Signed out, the header button is Start free trial and goes straight to starting one. Creating a free account to browse is still there, on the landing page.
- Until you track your first brand, an Add your first brand button sits in the header on every page, so wandering into Settings is no longer a dead end.
- Starting a trial or a new plan lands you on Add a brand, rather than on the billing page you have just finished with. Changing plan when you already track brands still returns to Plan & billing.
- If a renewal payment fails, every page now says so and offers to update the card. Checks carry on while the bank retries, rather than stopping without warning.
Alerts you choose, brand by brand
- Alerts now cover far more than a score move: a rival passing you, share of voice shifting, dropping out of an engine or being named by one for the first time, sentiment turning, a new negative theme taking hold, and the sites AI cites about you coming or going. Accuracy problems and Flare moves still arrive as they always did.
- Profile > Alerts lists every brand you watch, each set on its own: everything, important only, a weekly summary, or nothing. Behind each brand you can switch off a whole family of alerts — accuracy, competitors, or how the brand is described. A row at the top decides what a brand you add later starts on.
- Watching several brands, you can have one email each or a single email covering all of them.
- Every alert email is drawn in the brand's own colour, says what moved and how far, and opens straight at the tab that explains it. You can send yourself a test of any brand's alert from the settings page.
- Every alert email now carries an unsubscribe link that works without signing in, and can mute one brand rather than all of them.
- Starter now includes a weekly summary of where your brand stands and what moved. Alerts as each check finds them stay on Pro and Agency.
- The same alert no longer arrives twice when a check is retried, and a problem that persists waits before mailing you again.
Share of voice counts the competitors you listed
- Share of voice and average position measure a brand against the competitors listed on it, on questions that don't name the brand. Public profiles in the index were counting every other company an answer happened to name, which made the same brand look weaker there than in a workspace. Both now report the same number.
- Rival names written with a note in brackets, such as FAB (First Abu Dhabi Bank), now match the plain name an answer uses, so those mentions count instead of being missed.
Header images across the brand index
- Hundreds of brands in the public index now carry their own header image, along with the description, colours and social links that were missing with it.
- A brand whose only wide image is really a logo file shows the plain gradient band instead of stretching it.
- Where there is no photograph at all, the header is drawn in the brand's own colour with its initial set large behind it, so a profile looks finished rather than empty.
Owned, earned, social and paid on Sources
- Sources opens with the media mix behind the answers: how much of what AI reads about a brand is its own site, coverage it has won, communities and reviews, or marketplaces and listings. Each shows its share, its citations, how many of those landed on answers naming the brand, and its biggest sites.
- Every row of the websites table is labelled with its type, so you can see at a glance which cited sites are yours.
- Owned means your own website and anything under it. Social and paid come from lists we maintain, so a site we haven't classified counts as earned.
Archive a brand
- Settings has a Brands tab listing what you track. Archiving a brand stops its checks that day and takes it off your dashboard, while everything it has measured stays readable at its own page, marked Archived.
- Restoring brings a brand back exactly as it was, with its questions, fact sheet and history. Adding a brand you've had before restores it rather than researching it again.
- An archived brand keeps its place on your plan for 30 days, so brands can't be swapped for a fresh report every few days. Plan & billing says when the place frees.
- New brands are limited to three times your plan's brand count in a rolling year: 3 on Starter, 9 on Pro, 30 on Agency. Restoring a brand never counts.
A free plan
- New accounts start on Free: browse Explore and open any brand's profile, with its Flare score, visibility by engine and category ranking. Free accounts land on Explore, and an Upgrade to add brands button in the header and a note in the sidebar lead to the plans.
- Brand profiles now need an account. Opening one while signed out goes to a short sign-up, then straight back to the brand.
- Free is listed on Pricing and in Plan & billing alongside Starter, Pro and Agency.
- Get started and Sign up now create a free account straight away, without naming a brand first. Choosing a paid plan on Pricing while signed out creates your account, then goes on to payment.
- The free audit is gone. Looking up a brand in the index and opening its profile with a free account replaces it, and the guides and blog point there instead.
- Sign in and sign up are redesigned: the brand on one side of the screen and the form on the other, with the title in the brand gradient and the same fields and buttons as the rest of the product. On a phone the form takes the whole screen.
Clearer brand pages
- Brand pages lead with the brand itself: a one-line description, no repeated title above it.
- The Flare score card has a See breakdown link that opens Your scores: the five factors behind the number (Visibility, Prominence, Accuracy, Consistency, Sentiment), each with a question mark you can hover to see what it measures.
- The five headline numbers on a brand's overview are redesigned and renamed in plain English: Visibility, Share of voice, Average position, Accuracy issues and All engines. Each has a question mark explaining it, a bar showing where it stands, a sentence saying what it means (such as "Usually named 2nd" or "80 high · 15 medium · 5 low"), and a green or red change since 14 days ago.
- The rest of a brand's overview is reorganised so each thing is said once. What to do next ranks false claims, missed questions, negative stories and recommendations with a button for each; By engine shows how often every engine names the brand, including engines not checked yet; Trends puts visibility, share of voice and position in one chart with tabs, and says when there aren't enough checks to draw a line yet; Who else gets named ranks competitors by share of mentions, matching Share of voice.
- You can give a brand your own banner and logo from its overview: select the pen on the banner to edit, then Change banner or the camera on the logo, with the recommended size shown beside each. The change applies only to your workspace's copy of the brand, and Reset brings back the default. The logo also shows in the brand switcher and on your dashboard.
- While a brand's first check is still running, its page says so and counts the answers in so far, instead of showing empty numbers. If a check is interrupted before its scores are worked out, opening the brand finishes the job.
- Adding a brand now fills in Recommendations and Brand Narrative as soon as the first check finishes, instead of leaving them empty until the weekly refresh.
- Recommendations are redesigned: To do, Done and Dismissed tabs, recommendations listed most important first, and each one shown with its type (Content, Sources, Positioning, Technical), a one-line reason and numbered steps.
- Questions & Prompts is redesigned: filter by All, Not always named, Named everywhere or Paused; each engine's column shows 1st, 2nd and so on, or Not named, with a key above the table; each question says how often it's asked and what kind of question it is; Named by shows how many engines name you; Pause is easy to find; the tab row shows how many questions your plan allows per brand; and Add questions takes several at once, one per line, with a live count of how many fit your plan and repeats skipped.
- Competitors is redesigned: a ranked list of who AI names, with logos; a share-over-time chart where every brand has its own colour and a sentence on the gap to your nearest rival; and a head-to-head table in plain words, such as They usually lead or Neck and neck, with who's named first, each competitor's usual place and how AI describes them.
- Answer Accuracy is redesigned: Claims to review and Fact sheet are now tabs on the page. Filter claims by severity with coloured pills; each claim shows who said it, why it's wrong and what's true, with Not a mistake or Confirm mistake. The fact sheet can be filtered by To review, Verified or Rejected, shows each fact's source and how many answers contradict it, and lets you verify everything at once or add several facts, one per line.
- Brand Narrative is redesigned: filter themes by All, Positive, Neutral or Negative; each theme shows its tone, how often it comes up with a bar, what it says, and quotes from AI answers side by side with the engine that said them. Themes update on their own after new check results, so there's no refresh button.
- Sources is redesigned with two tabs, Most cited and Outreach opportunities. Each website shows its logo, how many times AI cited it, how many of those answers named you, and which engines cited it. Outreach splits into sites that are cited often but rarely name you, and sites that never do. Watched pages has been removed.
- AI Crawlers is redesigned in plain English. Until your site is connected, the page explains what it needs, in three steps, with a button to ask us for help and a folded section of details for your web team. Once visits arrive, it shows a summary of visits that answer questions now versus visits that train future models, a daily chart by company, each crawler named by its assistant with what its visits mean, and the pages AI reads most.
- Every brand page has a clearer top: the brand's update status shows as small chips, each section opens with its name as a heading, one sentence on what it's for and short notes on how it's measured, and the page lines up with the brand name.
- On the Overview, the bars on Visibility, Share of voice, Average position and All engines are coloured by how strong the result is, on the same scale as the Flare score. Share of voice now shows your real place among the brands named, where it always said most mentioned before.
- Opening an answer from Questions & Prompts shows it properly formatted, with headings, bold text, lists and tables instead of symbols, your brand highlighted and competitors underlined. Beside it are the brands the answer named, in order, with how it described them, and the websites it used.
- Loading screens match the new page layouts, the By engine bars on the Overview are coloured by strength, and an early Flare score is labelled 'Early score, still settling'.
- Main buttons have an orange-to-red gradient, taken from the brandflare flame.
- Team shows each member's profile photo and first and last name more reliably, including people added before they first signed in, and a letter instead of a placeholder picture for anyone without a photo. What each role can do uses clear green ticks.
- Plan & billing is redesigned: your plan with its price, renewal date and card; brands and seats used against your limits; plans side by side with a monthly or yearly choice; and recent invoices to download. Upgrading happens right there, with the amount due today shown before you confirm. To downgrade or cancel, send a request to the brandflare team from the same page.
- Alerts are now a personal setting: each person switches alert emails on or off for themselves in Profile, and they go to their own email. Alerts is no longer a workspace settings tab, and anyone who had alerts on keeps getting them.
- Logos that were white on a transparent background, and so looked like empty circles, now fall back to another logo or the first letter.
- The Dashboard is redesigned as a grid of brand cards. Each shows the Flare score in a square coloured by strength, then Visibility, Share of voice (with the brand's place among the brands named), Average position and Accuracy with coloured bars and changes, and how many accuracy issues are waiting. The sections below the grid have been removed.
- Public brand profiles are redesigned: the brand and its Flare score side by side, visibility for each engine with coloured bars, the category ranking beside it, and clearer previews of what tracking adds.
- Explore is redesigned: popular categories on the left with their average Flare score, and the week's biggest movers on the right. Can't find a brand? opens a short form in a pop-up, and search waits until you pause typing.
- When you're not in a brand, the sidebar lists your brands with their accuracy issues, and selecting your workspace goes to the Dashboard.
- Team uses the app's own dropdowns for roles, Add a brand explains clearly when you can't add one, and the page-not-found and error pages match the rest of the app.
- Pages fit better on a phone: your avatar sits beside the page title, section intros and recommendations use the full width, and long claims, questions and answers no longer run off the screen.
- The home page shows the new brand page design, Explore opens with a one-line introduction, and Learn and the blog are easier to read, with larger text, clearer headings and a card at the end of each article inviting you to create a free account.
- The sidebar's help and legal links sit in a tidy grid.
- The product name is written as lowercase brandflare everywhere: page titles, the home page, Learn, the legal pages, emails and share images, and it shows in the brand orange in page text.
- Public brand profiles look like a brand in your workspace: the banner, Flare score and the same sections, with everything but the Overview locked. Opening a locked section, or Add to your workspace, shows what adding the brand gives you and lets you add it after a confirmation, upgrade if your plan is full, or sign up first.
Teams, a fifth engine, and better reports
- Workspace admins can now invite colleagues from Settings › Team, give each person a role, and change or remove them. Admins add brands and manage people, billing and settings; editors change the questions, facts and actions on brands already there; viewers see everything and change nothing. Starter includes one seat, Pro three, Agency ten.
- Invitations arrive as a brandflare email that says who invited you, into which workspace and with what role. The link works for 30 days.
- A workspace can have a name and an icon, set by an owner or admin under Settings. Both sit at the top of the navigation, so it is always clear whose workspace you are in.
- A new layout: a sidebar on the left shows your workspace, a brand switcher and the current brand's sections with their open counts. Dashboard and Explore More Brands sit in the title row at the top of every page, beside the page's main action and your avatar. On a phone the sidebar opens from a menu button.
- Text is larger and easier to read throughout the app, and cards, buttons and forms share one consistent style.
- Settings is split into tabs: Workspace, Team, Plan & billing, and Alerts. Sources, Answer Accuracy and the admin catalog also use tabs, so each view has its own link.
- The avatar menu starts with your workspace, then your email and Manage profile, then Team, Plan & billing for admins, Alerts and Workspace settings, then Guides and Customer support, and Sign out at the bottom.
- Manage profile opens a Profile page laid out like Settings. Personal holds your photo, name and sign-in email. Security lets you set or change a password, shows your connected accounts, and lists signed-in devices so you can sign any of them out.
- Adding a brand follows the new layout, with three steps. Before research starts you confirm the brand name and website, since they can't be changed later. Setup shows live progress, and a Ready step sums up what was found before you open the brand.
- A brand's page opens with a welcome that explains the first check is running and usually takes 20–30 minutes, and what each section will show.
- If you leave while a brand is being set up, it stays easy to find. The brand is labelled Setting up in the brand switcher and on the dashboard, and its page shows live progress with a link back to setup. If research fails, the page says so and offers Try again.
- The client report link is off brand pages for now while client reports are redesigned. Report links you have already shared keep working.
- The brand page's Flare card shows a line for the last 90 days, and the status line under the description says when the brand was last updated and when the next update is due.
- Free search has been retired. Looking a brand up in the public index and the questions tracked on your own brands cover what it was for.
- Flare score bands are now Very high, High, Moderate, Low and Very low, and the score carries a short explanation on hover.
- Google Search joins ChatGPT, Gemini, Perplexity and Claude as a fifth engine. It is Gemini with live Google Search grounding, so it stands in for what Google's own AI answers say, and it shows its sources.
- Client reports open with where the brand is winning and losing, rank the questions where it should appear and doesn't, show which sites each answer cites, and split recommendations into quick wins and longer-term moves.
- The sweep alert email was rebuilt: what moved, by engine, with the top new accuracy flag and what is actually true.
- Every plan is now checked weekly.
Sibling brands and bigger comparison sets
- The judge can be told which similarly named organisations are not the brand, so a free zone, the district it sits in and the energy company that shares its name are no longer credited to each other.
- An account can be given more competitors and questions than its plan lists, for clients comparing a long field.
- An account set up on a client's behalf can be given a sign-in later.
Dashboards that say what the answers say
- The portfolio dashboard leads with findings: the false claims engines are repeating, the questions a rival is taking, and the actions to take, each named and clickable.
- The brand dashboard puts the Flare score in the header and makes every finding a link to the answers behind it.
- Answer Accuracy is a one-at-a-time review queue: each flag shows what the engine said and what is true, with acknowledge and dismiss, and a way to clear a backlog in one press.
- One sign-in works across environments, and viewing an account as an operator is strictly read-only.
Launch hardening
- Forty-six fixes from a pre-launch review: scoring corrections, new data views, and the essentials for going live.
- A public brand page explains why its score can differ from the tracked one.
- Adding a brand from the catalog copies its measurement history rather than asking every engine again, so a new account starts with a trend instead of a flat line.
- Checks resume where they stopped instead of starting over, and never run twice for the same question on the same day.
- An operator screen for accounts: see, create, edit and delete them, and open any one read-only.
The Flare score
- One number from 0 to 100 for how a brand stands in AI answers, built from five pillars: presence, prominence, accuracy, reach and sentiment. It replaces the raw visibility percentage as the headline figure and comes with a confidence level based on how many answers were scored.
A new shell
- Top navigation and a vertical brand rail, so which brand you are looking at and which tab you are on are always visible.
- A richer portfolio view with a sparkline and the latest score per brand.
The public index, sign-up, and the knowledge base
- Explore: a category-first directory of more than a thousand brands with a public score for each, browsable without an account.
- Brand-first sign-up: name your brand, create your account, and research starts while you watch. A guided setup and a short tour follow.
- No free tier and no trial: accounts are read-only until a plan is chosen.
- A knowledge base at /learn with guides, documentation and a glossary for generative engine optimisation, and a blog.
- Lighter research for catalog brands so the index can grow; a brand is researched in full the moment someone tracks it.
- Every AI model now runs through one gateway, and each accuracy flag is double-checked before it is kept.
- A redesigned landing page with a live finder and a look at how the judge reads answers.
Accounts, plans and the marketplace
- Sign-in and workspaces. Every brand, question and answer belongs to one account.
- Starter, Pro and Agency plans with their brand, competitor and question allowances, checkout and billing.
- A free AI visibility audit anyone can run from the site, with a shareable scorecard.
- Alert emails when a check finds a new high-severity accuracy problem or a visibility drop.
- White-label client reports on a private link, for agencies.
- The public catalog: tracked brands are cloned from a shared index into each account's private portfolio.
brandflare ships
- Ask ChatGPT, Gemini, Perplexity and Claude the questions your customers ask, and have a judge read every answer for whether your brand is named, how early, how favourably, and whether what is said is true.
- A dashboard per brand: visibility, share of voice against competitors, answer accuracy, the narratives engines repeat, the sources they cite, and a prioritised list of what to publish next.
- Weekly checks, staggered across the week.