What we shipped, when we shipped it. No fluff, just features.
27 May 2026
BrandCrux composite now favours visibility over infrastructure. Topic Authority and the BrandCrux dashboard hero used to give site-health signals (content depth, GEO audit) almost as much weight as AI mentions or SERP placements, so a brand with a 90/100 GEO score and zero AI presence could still score middling. The composite is now weighted 30% AI, 25% SERP, 15% citations, 12% socials, with the remaining 18% across content depth, crawl coverage, and GEO infrastructure. A perfect GEO audit no longer compensates for being invisible in AI answers or the top-3 of search results.
SERP column reads as a 0-100 score. The per-topic SERP cell used to show a raw top-10 placement count (e.g. "645" on a strong topic), which collided with every other pillar's 0-100 scale. It now reads as a normalised score (10 top-10 placements caps at 100). Colour bands match the rest of the table: green ≥70, amber 40-70, red below.
Empty competitor cells now show 0 instead of a dash. When a competitor returned nothing on SERP, citations, socials, website, or AI for a topic, the table rendered "no value" cells that looked like missing data. They now show 0 explicitly, since 0 signal IS the signal. GEO stays dashed when there's no site-wide audit on file (genuinely-missing data, not a 0-score).
26 May 2026
Discovered Competitors now classifies industry match per domain. The Discovered list used to surface every brand that appeared alongside yours, even when the candidate was a hospital, a regulator, a news outlet, or an unrelated industry. Each newly surfaced domain is now classified for industry fit against your workspace categories. Out-of-scope domains are hidden by default with a one-click "Show them" toggle, pending classifications are labelled "Pending review", and you can override any verdict with "Mark as competitor" so a false negative ever stays hidden again.
25 May 2026
Every public page now ships a complete social-share card and SEO meta set. New shared composable emits title, description, canonical, robots, OG (with image, dimensions, alt, site name), Twitter card (summary_large_image, site handle), and per-page-type JSON-LD (WebPage, Article, FAQPage, Product, BreadcrumbList). Sitewide Organisation + WebSite JSON-LD lands on every route. Sitemap now includes cookies, help, and status pages.
Your own social activity now sits next to competitors on every Socials view. Every Socials scan pulls posts from the brand's own LinkedIn, X, Instagram, Facebook, YouTube, and TikTok accounts alongside every tracked competitor. The engagement leaderboard and Authority Topics scorecard pin "You" to row 1, the per-topic expansion shows your engagement against each competitor side-by-side, and a new "What to post next" panel produces 5-10 specific post ideas where competitors are outpacing you. New brand-self handles card at the top of the page handles add, remove, and AI-suggest with one click. Onboarding auto-suggests handles when you create a business, so the first scan has full self vs competitor data without any manual setup.
Mission Control flowchart cards stop chopping words mid-character. Long-form tier names and summaries ("Marketcheck Cars Inc marketcheck.com Automotive Data") used to wrap one letter per line on narrow viewports. Tightened the wrapping rule to break on whitespace only, and titles now span up to 2 lines instead of ellipsis-truncating.
AI Visibility insight row is balanced again. The four cards (Personas, Top citations, Competitors, Insights) sit in a 2x2 layout instead of leaving Insights orphaned in row 2.
Web Content Pages table no longer pushes columns off-screen. Long URLs in the Page column used to truncate the Words / Score / Last seen columns into invisibility. The grid layout now reserves min-widths so every column stays readable.
Help info on every Web Content tab. Each tab (Overview, Pages, Topics, Gaps) now has a collapsible help section explaining what's shown, how the data is built, and what to do with it.
Web Search deep-dive opens as a closable tab. Clicking a keyword on the Web Search Queries page now opens a tab next to Overview / Queries / Pages, so you can keep multiple keyword deep dives open and switch between them without losing the listing. Close any tab to dismiss it.
Socials Off-topic section opens by default. The "what else competitors are posting about" view used to start collapsed. Now it expands automatically with the top 10 rows in view.
22 May 2026
Public pricing page — A new /pricing page lists every plan side-by-side with a monthly / annual toggle, top-up packs, a feature-matrix table, and pricing FAQs. Annual savings call out automatically when you flip the toggle.
Annual billing back for new sign-ups — Razorpay's per-charge limit was raised, so the monthly / annual toggle in the in-app billing panel works for new subscribers too. Annual saves up to 15% across Starter, Growth, and Scale.
Settings → Run the onboarding wizard — A new card in account Settings opens the guided onboarding flow on demand, so dormant users (or anyone who didn't finish the wizard on signup) can re-enter at any time without going through Mission Control or signup again.
Clearer GDPR + CCPA story — The Privacy Policy now spells out California Consumer Privacy Act (CCPA / CPRA) rights alongside the existing GDPR coverage, and the in-app "Delete my account" flow explicitly references the right-to-erasure laws under which it operates.
Admin nav now exposes Leaderboards + Social drafts — The content-marketing engine's surfaces (industry-leaderboard jobs, topic / persona / prompt libraries, social-post draft queue) now have entries in the admin sidebar instead of being reachable only by direct URL.
Cleaned up legacy comparison pages — The five per-competitor head-to-head landing pages were retired during the rebrand. The main /compare hub stays — same matrix, same decision tree, all five competitors in one place.
Every dashboard tile is now a 0-100 score that matches its channel page — SERP, Citations, and Web Content each render a vs-leader score (your count as a share of the leading competitor's, capped at 100) as the headline number on both the BrandCrux tile and the corresponding channel page. The supporting counts ("76 top-10s · leader has 100") sit underneath. AI Visibility and GEO Audit continue to show their native 0-100 scores. Socials renders "—" until brand-side posting tracking comes online. Click a tile, land on the page, see the same score.
Sharpened the Growth-tier story — The marketing page now leads with what's live today: AI Assistant, Content Engine briefs, CMS publishing across WordPress / Shopify / Webflow / Wix / Squarespace, REST API + MCP server, and on-demand Authority Cards. Tighter copy, no maybes.
21 May 2026
Guided onboarding wizard — every workspace lands fully wired — Adding a new business now walks through a five-step flow: brand basics → competitors (AI-suggested + manual, name and website required) → authority topics → tracked prompts per topic → confirm. Each step auto-generates suggestions on first arrival and lets you regenerate if you change earlier inputs. On confirm we kick off keyword discovery, backlinks + traffic refresh, social handle suggestions, and the GEO audit — and optionally fire your first AI Visibility scan. Dashboards start populating with real signal within minutes of finishing instead of after a manual sequence of channel clicks.
Authority Topics page now loads reliably for older workspaces — A regression earlier today caused the Authority Topics list to show as empty for workspaces that had topics defined, while the Topic Map below kept showing data. Both panels now render against the same source so the page is consistent.
BrandCrux dashboard polish — The KPI tile strip is now a balanced 4+3 layout instead of leaving a lonely tile alone on row 2. Tiles also share a common baseline across titles, scores, and footer descriptions, so the row scans cleanly even when one card is still waiting for data. The Share button now lives inside the Topic Authority hero card, top-right, instead of floating above it.
Mission Control: removed the noisy status strip — The strip above the flowchart polled the pipeline state every few seconds but rarely produced anything new. The flowchart and per-step drawers already surface every next action with a real CTA, so the strip was just visual noise above them.
Keyword Discovery — one click, every keyword you and your competitors rank for — A new Keyword Discovery page (linked from Mission Control as Tier 1d) pulls every keyword your brand ranks for, plus the same for every active competitor, with search volume, difficulty, AI Overview flag, and ranking URL. Discovery seeds the topic clusterer, the prompt-to-topic backfill, and the competitor-gap reports automatically — so brands that onboard without Search Console finally have something for the clusterer to work with. Adding a new competitor now also fires a discovery scan in the background, so authority topics light up on its first appearance. 15 credits per domain scanned.
Auto-discover competitor social handles on add — Adding a new competitor used to leave the socials surface blank until you opened the suggestions modal. Now we look for the handles upfront — pulling them from the crawl footer when one is available, and from a confident model lookup when it is not. High-confidence handles auto-attach so the Socials page lights up on first scan instead of after a second round-trip.
Suggest socials is more resilient — When the primary suggestion model is rate-limited or unavailable, Suggest with AI now retries through a second model instead of failing back to a generic error. Same cost ceiling, far fewer "couldn't generate suggestions" messages.
Dashboards always surface scan data — no more "No signal yet" after a successful scan — Fresh accounts without keyword coverage used to land in a silent gap where BrandCrux and Authority Topics showed empty cards even after a completed AI scan, because the per-topic aggregation needed a topic→keyword→prompt chain that didn't exist yet. We now auto-fire keyword discovery on business creation, on every newly-added authority topic, and again as a safety net when a scan completes without topic mapping. The topic-snapshot writer also writes a fallback distribution across your authority topics so the dashboards never sit empty after a successful scan.
Mapping readiness banner on Mission Control and the scan launcher — Before scans fire, we now check that authority topics, keywords, prompts, and topic-mapping all exist. If anything is missing, a banner surfaces the exact next step (Add topics / Run discovery / Add prompts / Map prompts) with a one-click action. No more spending credits on scans that can't aggregate.
Email when a scan finishes — Every scan type now sends one completion email — AI Visibility, Web Search, Citations, Socials, Search Console, GA4, and Competitor Crawl. The email has the report name, two or three headline numbers from the run, and a direct link to the report. Toggle it off in Settings under Notifications. The topbar "scans running" pill continues to work whether the email is on or off.
Authority Topics Scorecard is now full-width on the Socials page — The per-topic scorecard was previously sharing the row with the engagement leaderboard, which truncated topic names and clipped the per-channel breakdown. It now spans the full width of the page so every topic's posts, engagement, and top competitor are visible without squeezing. The Off-Topic panel also got a Show all / Show top 10 control so long lists no longer drown out the rest of the page.
Every channel is now shareable — Public shareable reports rolled out to every channel: Web Search, Directory Citations, Socials, Authority Topics, BrandCrux Summary, GEO Audit, and Search Console Insights. Each page has a Share button that generates a stable public link — re-clicking returns the same URL rather than fanning out duplicates. Each report is read-only with no edit affordances, sized for stakeholder skim rather than deep ops. Starter and above.
Search Console + GA4 scan triggers are now gated behind their connectors — The Launch a scan page used to let you queue a Search Console or GA4 deep scan even when no OAuth connection existed; the run would silently fail downstream. Those channel cards now show a dashed "Not connected" treatment with a direct "Connect Search Console →" / "Connect GA4 →" link so you set it up first.
Competitor "Extract from Crawl" socials actually finds handles now — A key-rename in the crawl pipeline left every competitor crawl writing an empty page body, so Extract-from-Crawl returned zero handles every time even when the footer clearly listed them. The crawler now reads the right field; re-crawling a competitor with social-link footers populates handles as expected. Suggest-with-AI also surfaces real error detail instead of a generic "failed" message when the model errors out.
"Scans running" indicator in the topbar — Whenever a scan is in flight for the active workspace — AI, Web Search, Citations, GSC/GA4, Crawl, GEO Audit — a small pulsing pill appears in the topbar showing the count (e.g. "3 scans running"). Tap it to jump straight to the Scans page. When all scans complete you also get a toast nudging you to view the output.
Add a business actually adds a business again — The "+ Add business" item in the workspace switcher used to redirect to a stub page that bounced back to Mission Control for the current workspace, leaving users with no way to create a new one. It now opens a real new-business form (name, website, category, products/services, "auto-fill from website") and drops you into Mission Control for the freshly created workspace.
Content brief drafts: honest progress instead of a premature "ready" toast — Tapping Generate Now on a brief used to flash "is ready for review" within seconds even though the writer was still composing. The success banner now says the draft is being generated, offers a direct "View draft →" link to monitor it, and the draft page itself shows a live "Generating your draft…" loader with auto-refresh — the editor populates the moment the writer finishes.
Persona scans now route through the unified scan launcher — The Personas page previously opened its own dedicated "Run persona scan" modal. It now links straight into the main Launch a scan page with the AI Visibility section already in view, so persona-scoped scans share the same channel picker, cost estimate, and approval flow as every other scan. One launcher, fewer surprises.
Mission Control cards render at full width again — A regression yesterday squeezed each tier card's title and summary into a narrow left column when the corner page-link arrow was added. Cards now fill their grid cell so titles like "About your brand" no longer truncate to "Abo…".
Adversarial watch cost is now clearly disclosed — The Competitors page tooltip for Adversarial watch previously quoted "about 1 credit per day". Updated to the accurate "about 50 credits per day for up to 50 pages crawled" so the cost of toggling it on per competitor is upfront. The feature is opt-in, off by default, and toggled per competitor.
GEO Audit listing repaired — The GEO Audit page was rendering an empty list for both your own audits and competitor audits, even when completed audits existed in the database. Fixed the underlying response shape so every audit you've ever run shows up again. Also patched the deep link used by the unified Scans history view so clicking a GEO Audit row from there opens the right page instead of a 404.
Mission Control cards now jump to where you configure them — Every tier card on Mission Control, in both the graph view and the new list view, is now a link to the matching configuration page (Authority Topics, Competitors, Connectors, Personas, Prompts, AI Pulse, Web Search, Socials, etc.). Title click in the list, corner arrow on the graph card — no more bouncing through the sidebar to act on what Mission Control is telling you.
Mission Control reads bigger — Bumped the whole page up 15% so the tier names, status pills, and progress numbers are easier to scan at desk-distance.
Anomaly + competitor-move cards on your dashboard are now actionable — Each row in the activity feed on the BrandCrux dashboard now navigates somewhere: anomalies open in Trend Intelligence, competitor moves open in Competitor Watch, and published content rows open the brief. No more cards that just sit there telling you something happened.
20 May 2026
Topic Setup Wizard — A new wizard for every authority topic. Click Setup wizard on any topic card and BrandCrux proposes 8 buyer-intent prompts plus 20 SERP keywords for that topic, surfaces any GSC queries that look related so you can pull them in with one click, and lets you edit or add your own before applying. Adding a new authority topic auto-opens the wizard so it's wired into the moment of intent. Apply writes the prompts and keywords atomically; the next pipeline run picks them up.
Topic leaderboards now show every tracked competitor — Previously a competitor only appeared on a topic leaderboard if a scan had returned data for them. They now appear on every topic, with — chips in the channel cells where no data exists yet and a compact Run pipeline → link on rows that have nothing at all. Clicking the link deep-links straight into the Run Pipeline modal on Competitors with that competitor pre-selected, so backfilling is one screen and one confirm. The data gap is now visible and actionable.
GEO score is now per-topic, not site-wide — The GEO chip on every topic card was previously the same site-level score copied across every topic. It now averages the actual GEO Audit scores of pages mapped to that topic (relevance-weighted), so a brand with great schema on their pricing pages but missing schema on their guides shows the difference. Topics with no page-level GEO data show — rather than a misleading number.
Run Pipeline modal — centering + checkbox bug fixed — The Run Pipeline modal on the Competitors page now centers in the viewport regardless of scroll position (it was getting trapped inside a stacking-context ancestor on long pages). Toggling a stage checkbox no longer removes the row from the list — the modal keeps every stage visible so you can change your mind. Total estimate updates only based on what's checked.
Web Search / Queries now shows positions — The Queries view was rendering Not ranked for every keyword regardless of actual data. Fixed the response-shape mismatch behind the page so real positions and competitor-rank columns populate again.
In-progress pages tucked out of production — Automations, Plays, and Attribution were appearing in the sidebar on live brandcrux.ai while still under active development. They now stay visible on localhost (so the team can dogfood) and hide from the production sidebar until they ship.
BrandCrux is now the Brand Authority Tracker across channels — Sharper category claim, same product. The homepage hero, every footer signature, page titles, social previews, and the public llms.txt now describe BrandCrux as the Brand Authority Tracker across channels — AI, Web, SERP, Socials, citations, and your own analytics. The seven-channel breakdown stays on every deep page, the Topic Authority composite stays the headline metric, the closed-loop measurement stays the differentiator. Only the words around them got tighter.
1,000 free credits on signup, fatter referral rewards — Every new account now starts with 1,000 free scan credits (up from 300) — enough to run a full scan across all seven channels for one business and see the composite light up before you decide on a paid plan. The Refer & Earn rewards are also bigger: 500 credits each (up from 100) when an invitee signs up, and 1,000 credits each (up from 500) when they upgrade to a paid plan. Existing balances stay intact; new grants apply to all signups from today onwards.
19 May 2026
Homepage rewrite — straight to the question that matters — the landing page hero now asks the actual question ("Your customers asked ChatGPT, Google, and Reddit. Did your brand show up?") and runs a free scan inline. Drop your domain, hit enter, get the truth. The Without / With comparison moved up the page so the value lands before the deeper sections, every channel card now leads with the outcome you get (not the mechanism behind it), and every step of the Loop carries a concrete example of what that step produces ("+12 pts the week after the FAQ schema shipped"). A slim sticky CTA now follows you down the page so the free scan is always one tap away.
FAQ now opens to substantive answers — the first two FAQs on the homepage and the AI Visibility page expand by default. Visitors see real answers without having to click first.
AI Visibility pricing comparison matches the live plans — the desktop pricing table on /ai-visibility previously showed the pre-rebrand credit pools (3,000 / 12,000 / 25,000) and a stale 100-credit Free tier. It now reflects the real ladder: Starter 10,000 credits, Growth 50,000, Scale 100,000, Free 300 one-time credits, plus team-member counts and an accurate REST API / MCP split.
Recent Scans hides the zero-score noise — the recent-scans strip on the AI Visibility page no longer shows scans whose overall score never came back above zero. Those rows were giving visitors the wrong read on the product. Indexed, scored businesses only.
Source citations on the Uncomfortable Math — every headline stat on the AI Visibility page ("100M+ daily AI searches", "73% trust") now carries a one-line source attribution underneath, so the numbers stop looking like vapor.
Bigger top-up packs to match the new credit appetite — Top-up packs rebuilt at clean $25 / $50 / $100 / $250 price points for 1,500 / 4,000 / 9,000 / 30,000 credits respectively. The old 500 / 1,500 / 3,500 / 6,000 ladder was sized for the previous monthly pools (3k / 12k / 25k) and felt cramped after the upgraded plans (10k / 50k / 100k). Top-up credits still never expire, and the per-credit cost is calibrated so a sustained-usage customer always gets a better rate by subscribing to the matching plan.
New pricing, in USD across the board — Plans rebalanced and now billed in USD on every checkout. Starter moves to $99 a month and gains REST API access plus the AI Assistant (previously gated to Growth) and 10,000 credits a month. Growth moves to $299 with 50,000 credits, MCP server, content engine, and CMS publishing. Scale moves to $499 with 100,000 credits, white-label reports, priority support, and a dedicated CSM. Annual billing offers a 15% discount across every paid tier. Scale is now visible again in the plan picker after a temporary pause. Top-up credit packs unchanged.
Citation gap list no longer flags a competitor's own properties as "gaps" — the gap list on Directory Citations and on every topic page now filters out citations on the competitor's own website or their own LinkedIn / Facebook / X profile. Only third-party publisher and directory citations remain — those are the actual gaps worth closing.
Web Search competitor leaderboard skips non-competitor platforms — the leaderboard on Web Search no longer surfaces social platforms (Facebook, Instagram, LinkedIn), reference sites (Wikipedia), big news outlets, consulting firms, code hosts, or app stores as "competitors". The Domains Observed KPI still counts every root domain seen in top-10 — it just doesn't pretend Deloitte is a competitor.
Engagement & geography backfills — the bounce rate, pages-per-visit, avg duration, and top-country fields on Backlinks & Traffic now populate for every tracked competitor after the next traffic refresh. Existing rows can be backfilled in place without re-pulling the rest of the snapshot.
Topic Map now reflects competitive reality — major competitors get their full property rollups crawled (Cox Automotive's Manheim / vAuto / KBB / Autotrader / Dealertrack, J.D. Power's NADAGuides / Autodata, Edmunds' CarCode, Black Book's blackbook.com, S&P's IHS Markit). Crawl page budgets scale per-competitor (Cox 400, J.D. Power / S&P / Edmunds 300, the rest 200). Hard-walled sites that used to silently 0-page now route via proxy after an SSL fix so the bot-walled retries actually work.
Keyword detail page — clicking any keyword on Web Search now opens a per-keyword detail view with position-over-time sparkline, latest top-10 SERP, per-competitor board (best position each domain has held), and People-Also-Ask. The old click-into-404 is gone.
Pages tab shows every ranking URL — Web Search / Pages previously only returned featured-snippet rows (most businesses have zero), so the page looked empty. It now shows every URL that turned up in SERP for your tracked keywords — organic, People-Also-Ask, and featured snippets together.
Backlinks & Traffic upgrade — Competitor Intelligence is now titled Backlinks & Traffic to match the Channels card. Every table (backlinks, traffic, keyword gaps, new + lost links) gets click-to-sort column headers. Two new widgets: DoFollow link quality with a strength bar, and Engagement & geography (bounce rate, pages per visit, average visit duration, top country) for every tracked domain. A "How to read this" panel explains each widget.
Competitor social handles, three ways — open any row on Competitors and tap "Socials" to manage handles. Add manually, ask AI to suggest the brand's known accounts, or extract them from the footer of an already-crawled site. Every completed competitor crawl auto-extracts LinkedIn, X, Instagram, Facebook, YouTube and TikTok URLs going forward.
New scan: GEO Audit channel — the new-scan picker on Start a scan gains a "GEO Audit" channel. Toggle Self and/or Competitors, set the max-pages slider, and the run dispatches an audit per target. GEO audits show as their own rows on the Scans page with status, score, target domain in the params chip.
Socials pre-flight — selecting the Socials channel on Start a scan now warns when no competitor has any social handle configured (which would have run a zero-result scan), with a one-click path to Competitors to fix it.
Personas: Suggest with AI fixed — the Personas page's Suggest button previously returned a generic error for some brands because the LLM's output sometimes contained bare numeric ranges that strict JSON parsing rejected. Now repairs malformed JSON automatically and generates 4 to 6 buyer personas reliably.
Bug sweep — Channels page tiles for Citations, GSC Insights, GA4 Insights showed empty values for businesses that had data (column-name mismatches in the underlying queries) — now populate. Authority Topics cards get visible borders so individual topics no longer blend together. Directory Citations queries cap explicitly at 1000 rows to avoid silent truncation.
18 May 2026
Topic Authority now scores each topic distinctly — every priority topic gets its own composite score based on the actual content depth + relevance of the pages covering it, instead of echoing one site-wide audit number. Multi-topic pages count toward each topic they cover.
New: Content Authority — filterable per-page score table covering your site and every tracked competitor. Filter by topic, by source (you vs them), sort by authority / technical / freshness / relevance. Click any page for the 90-day score history per topic and the full signal breakdown.
Two grading modes you can pick per scan — at the start of a Discovery run, choose Technical-only (no extra cost on top of the crawl — structural signals, schema, freshness, link depth) or Technical + Content (+5 cr per page for AI-graded depth + per-topic authority). Your last pick is remembered.
Shared page-score index — Force-fresh override — when a competitor URL is already in your platform's index and fresh, the crawl step is automatic. If you know a competitor just shipped something significant, tick "Force fresh crawl" to skip the index and re-pull live.
Board Report Section 7 — the report now includes top brand pages per priority topic alongside competitor top pages, with the authority score and grading mode for each.
New scan picker shows Web GUI engines — the engine list on Start a scan is now split into two tiers: API tier (direct model API, 1 credit per cell) and Web GUI tier (the real product UI scraped in real time, 5 credits per cell — ChatGPT, Gemini, Perplexity, Copilot, Grok, DeepSeek, and Google AI Overview). Mix the two to balance cost against "what would a real user actually see right now".
Live progress on the Scans page — every running scan on Scans now shows a pulsing status pill, a live elapsed timer, and a thin progress shimmer along the row. The page auto-refreshes every 20 seconds while anything is in flight, so a Running scan flips to Completed or Failed on its own. After kicking off a new scan, the launch panel links straight to the Running tab.
Backlinks & Traffic now on Channels — the Channels hub gains a Backlinks & Traffic card showing your brand's referring-domain count, the 7-day delta, and a 14-snapshot sparkline. Click through to the full Competitor Intelligence board.
Competitor crawls fail honestly — when every page request gets blocked and zero pages come back, the scan now reports as Failed (not Completed). When more than one competitor is requested in the same kick-off, all of them get dispatched instead of just the first. The crawler also retries via proxy on a hard block instead of giving up after a single attempt.
Each scan row shows its parameters — every row on Scans gets a small "params" chip. Hover or focus to see exactly which engines, country, page budget, depth, etc. the scan ran with. Helps when comparing two runs of the same channel side by side.
Crawl-time + AI-suggested competitor socials — finished competitor crawls now extract LinkedIn, X, Instagram, Facebook, YouTube and TikTok handles from the page HTML automatically, ready for the next Socials scan. When the crawl came up empty, a new "Suggest with AI" path proposes the brand's known handles for the user to confirm.
Assistant history is easier to find — on Assistant, empty "New conversation" stubs are hidden, and the mobile toggle now reads "History · N" with a badge so past chats are obvious.
Personas: AI suggestion fixed — Suggest with AI on Personas previously returned a generic error for some brands; now generates 4 to 6 buyer personas reliably.
17 May 2026
Competitor Intelligence — backlinks, traffic, keyword gaps — A new channel arrives behind your login. For your brand and every tracked competitor: total referring domains and domain rank, estimated monthly organic visits with month-over-month change, and the keyword gap list ranked by opportunity score (search volume × ease × competitor rank tier). Open it at Competitor Intelligence.
New + lost backlinks feed — Every week, the system pulls fresh backlink data and surfaces what changed: new referring domains gained, links lost (with the referring domain rank so reclaim outreach is prioritised). The feed lives inside Competitor Intelligence and rolls into the weekly admin email so you don't have to check.
Board Report adds a Backlinks · Traffic · Keyword Gaps section — Section 6 of every Board Report now includes the full competitive board: brand vs each competitor on referring domains, monthly visits with delta, the top 15 keyword gaps by opportunity, and the latest 30 days of new/lost links.
Built for cost — Every data point is cached with a TTL (backlinks 7 days, keyword portfolio 14, traffic 30). The "Refresh now" button is rate-limited to once per business per 24 hours. The weekly auto-refresh respects the cache, so it costs ~5 credits per business on a steady-state week.
16 May 2026
One brand, one home — The product you knew as LLMSight is now BrandCrux, live at brandcrux.ai. Every bookmark, link, and email click-through still works — older addresses redirect to the same page on the new domain. Your login, scans, credits, billing, API keys, and team members carry over unchanged.
15 May 2026
Board Report landing page rebuilt — A clearer first impression: a copy-led hero with a centred "Starting at $99" gradient block, then eight alternating feature rows. Each row pairs benefit-led copy with an actual widget preview pulled from the product — composite scorecard, Topic Authority leaderboard, AI engine sources grid, diagnosis feed, 90-day Plays board, citation gap list, competitor tracking grid and CMO memo. The sign-in card is now a full-width "Start your Board Report" panel with no step indicator on first view, so visitors aren't greeted with a nine-step tunnel before they've decided to start.
Email sign-in switches to a 6-digit code — Instead of waiting for a magic link to land and bouncing back to a new tab, the Board Report sign-in now sends a numeric one-time code that you can paste straight into the page. Continue with Google is still the one-click option above it.
Delivery window: 24-48 hours — Board Report turnaround tightened from "within 72 hours" to "24 to 48 hours" across the landing page, order confirmation, and emails.
New: $99 self-serve Board Report flow — The Board Report is now a guided product, not an email handoff. Sign in, walk through brand → competitors → topics → optional Search Console → prompts → engines → channels → pay. Each step persists as you go. The first 5 competitors, 5 AI engines, and 15 prompts are included in the $99 base; meters on every step show what you're choosing and any overages. Final order summary itemizes the breakdown and you can jump back to any step to tweak. A real BrandCrux account is provisioned at sign-in and activated with 1,000 credits on payment. The report is delivered in your account and emailed within 24 to 48 hours of admin approval.
Coupons + targeted invites — Admins can issue percent-off or flat-off coupons, global or scoped to a single email, with optional expiry + usage caps. 100%-off coupons skip the payment step entirely. Invite composer sends a personalized email to one or more recipients with the coupon attached and the recipient's brand context prefilled into the flow.
Monthly nudge — Customers with a Board Report older than 30 days get a one-line monthly reminder with a prefilled link to order a fresh one.
14 May 2026
(superseded — see 15 May 2026 entry) The BrandCrux Board Report, $499 one-time — A comprehensive, board-room-ready assessment of your brand's visibility across every channel BrandCrux tracks, delivered as a signed ten-section PDF in 72 hours. Composite Topic Authority with a category baseline. Pillar contribution. Competitive battlefield, topic by topic, every named competitor benchmarked. Typed diagnoses ranked critical / warning / info. A 90-day Plays roadmap with forecasted lift and verification signal per Play. Citation gap list. Owned-channel synthesis from Search Console and GA4 written by a senior analyst. GEO scorecard. Flywheel state. And a three-page CMO memo that ties everything together. Every order includes 14 days of live workspace access, a 30-minute walkthrough call with the analyst who built it, and a full refund any time before delivery. No subscription required. Place an order at /brandcrux-report.
12 May 2026
Smarter credit pricing — Credit cost per call now reflects the actual model your scan or content brief runs on. Heavy reasoning models cost a few more credits per call, light models stay the same, and free models stay free. Total credit usage for a typical scan is unchanged or slightly better — only the model-mix shifts the bill. The credit approval card always shows the upfront estimate before you run anything.
Automations — Watchers, Playbooks, and one-click push to Slack and Linear — A new Automations entry in the sidebar opens a three-tab control plane. Watchers are if/then rules: pick a signal (competitor content burst, composite score drop, topic share drop, search rank drop, citation went stale), set a threshold and a time window, and pick an action (spawn a Play, post to Slack, file a Linear issue, or send an email). Hit Test on any rule to dry-run it against current data without firing. Playbooks are pre-built multi-step sequences — two come stocked: Close a citation gap on a top directory drafts the listing copy, opens it for review, waits for the citation to appear, then measures the lift; Counter a competitor content burst spawns a counter-content Play, drafts the brief, posts to your team channel, and measures the lift. Run any Playbook on demand, or let it auto-fire when its trigger diagnosis lands. Destinations connects Slack via Incoming Webhook URL or Linear via Personal API Token — credentials encrypted at rest, never returned to the browser. Every Play detail page gets a Push to... button that drops the Play title, hypothesis, forecast, effort, confidence, verification signal, and a deep-link into a Slack message or Linear issue with one click.
One Channels hub replaces seven separate sidebar entries — AI Visibility, Web Search, Web Content, Citations, Socials, GSC Insights, and GA4 Insights have collapsed into a single Channels entry. Clicking it opens a unified channel grid where each tile shows the channel's current KPI, last activity, and a 14-day trend sparkline. The legacy URLs all still work and every channel name still appears in the command palette, so you can keep typing "GSC" or "Web Search" to jump straight in. Seven decisions a user shouldn't have had to make in the first place have become one.
In-product schema editor lives inside every schema-verified Play — Open any Play whose verification signal is "Schema present" and a side-by-side schema editor appears under Generate schema. Pick from ten schema types (Product, FAQ, Organization, Article, VideoObject, HowTo, Review, LocalBusiness, Service, BreadcrumbList), fill in the required fields with live validation, and watch the JSON-LD render in real time on the right. Copy it, download it as a .jsonld file or a ready-to-paste <script> tag, or click Apply to Play to save the block to the verification target so the verifier compares the live page against the exact JSON-LD you drafted. Schema is no longer "ask the LLM and paste blindly" — it's drafted, validated, and bound to the Play that proves it shipped.
Content briefs can be generated straight from a Play — When a Play's verification signal is "Content published", the detail page now exposes a Generate brief from this Play shortcut that drops you into Content Hub with the topic, hypothesis, and target URL pre-filled. One click instead of three.
Plays close the loop on every recommendation — A new Plays page in the Act section consolidates your atomic recommendations into outcome-shaped units of work. Each Play rolls up one or more recommendations under a single hypothesis with a forecasted visibility lift, and binds a verification signal that proves the work shipped: a published URL with measured lift, a structured-data block appearing on a target page, a citation showing up on a target domain, or a sustained click delta on a target query. No more "mark done" on the honor system. Ship a Play and an automated worker keeps polling the chosen signal until evidence lands — only then does the Play flip to Verified, and the realized lift is captured next to the forecast so your forecast accuracy becomes a moat over time. A board view ranks open Plays by priority, a detail view exposes the bound recommendations, content briefs, verification evidence, and attribution edges, and a one-click Consolidate rolls active recommendations into the right Plays with the right verify signal pre-wired.
Design system pass: consistent dates, statuses, and self-row markers across every page — Recommendation categories now read as Title Case ("Structured Data", "GA4 Deep") instead of all-caps slugs. Every status indicator across Connectors, Scan History, Mission Control, and AI Citations renders through a single pill component with one consistent set of variants (success / warn / error / neutral) so "Connected", "Synced", "Partial", and "Errored" all look like they belong to the same product. Every leaderboard row that points back at your own brand now uses the same green-dot "You" badge — no more mix of stars, solid pills, or coloured dots. Dates auto-flip from relative ("2h ago", "Yesterday") to absolute ("11 May") at the 7-day mark, and "Select a business" empty states now share one short sentence plus a "Or create one" link wherever they appear. Persona KPI cards offer a "Run your first persona scan" call-to-action inside the empty cards instead of showing a bare em-dash, the Web Search SERP leaderboard adds a position-distribution column so it's obvious how a competitor's appearances split between page 1 and the long tail, and engagement-per-post numbers keep one decimal place instead of rounding 2.4 down to 2.
Recommendations and content briefs no longer pile up duplicates — Near-duplicate recommendations (five "Implement Comprehensive Schema Markup" variants, six "vs competitor" pages) now collapse into a single canonical row with a small "+N similar" badge you can click to see what was rolled up. The headline "estimated visibility points available" uses a diminishing-returns curve calibrated against your current score, so it stops claiming impossible numbers like 345/100. Content briefs are deduped the same way: at generation time the advisor skips any candidate that fuzzy-matches a brief you already have, and the briefs list shows one canonical per idea.
Content Hub now owns the full Brief → Draft → Published → Lift loop — The separate Content Actions page has been folded in as the new Lift tab on Content. Every brief gets a Mark as published button that captures a baseline visibility score, schedules a 14-day post-scan, and surfaces the measured lift in the Lift tab once the post-scan completes. Old links to /app/content-actions redirect to the Lift tab so nothing breaks.
Entity names are now deep-linked across every report — When a recommendation mentions a competitor, topic, or keyword you already track, the name becomes an inline link straight to that entity's canonical page. Authority Topics competitor rows + topic-map columns now open the matching competitor page. The Discovered Competitors panel drills into the channel that surfaced each candidate (SERP, citations, or AI mentions). Top citations on AI Visibility deep-link by domain to either the competitor row (when tracked) or the citation drill page. Scan History rows now open AI Visibility focused on that scan. The "closed loop" from a recommendation to the prompt, scan, or competitor that triggered it is now a single click.
Intuitive URLs now work everywhere — Typing /app/ai-visibility, /app/gsc, /app/ga4, /app/citations, /app/web, or /app/topics now lands on the right page instead of a 404. Bookmarks and shared links survive even when the canonical route uses a longer slug.
Consistent browser-tab titles — Every app page now reads "Page Name · BrandCrux" in the tab. Replaces the mix of —, |, and missing-title fallbacks that crept in over the last few sprints.
New Scan button shows a credit estimate at a glance — The topbar New Scan button now reads "New scan · ~Ncr" based on your active prompts and the default 3-engine AI run. Hover for the formula. The launch page still recosts precisely once you pick channels and tune parameters.
Sidebar scan-status indicator is no longer mysterious — The small clock next to AI Visibility is a scan-status badge, never a deprecation marker. Hover now explains exactly what it means ("A scan is currently running for this business" or "Auto-scan scheduled — daily at 09:00 UTC").
Connector failures now tell you what went wrong — A failed Google Analytics or Search Console sync used to show a red "Errored" pill with no detail. Now every errored connector surfaces a plain-language hint ("Google Analytics didn't respond — try again or reconnect", "Access has expired — click Reconnect"), a "Show details" toggle that reveals the raw error for support, an "Errored 2h ago" timestamp, a one-click Retry sync button, and a Reconnect button that re-runs authorisation while keeping your historical data. Mission Control's "Sync owned analytics" step now also turns red when any underlying connector is errored — so the dashboard stops showing a green tick while a source is silently broken.
Duplicate workspaces no longer count against your limit — If two workspaces share the same name and domain (a common side-effect of running setup twice), only one counts toward your plan's business limit. The workspace switcher now picks the most recently updated row as canonical and the dropdown ends with an always-visible "n / m workspaces used" footer so you know exactly where you stand against your plan's cap. At the cap, the footer adds a tier-aware upgrade prompt (e.g. "Upgrade to Scale (5 → 25)") and the + Add business row goes muted with a tooltip explaining why. Workspaces never get deleted automatically — the duplicates simply stop double-counting.
Brand names no longer split across pages — "Carfax" and "CARFAX", "Marketcheck" and "MarketCheck", "Black Book" and "blackbook.com" now resolve to the same row everywhere. A new canonical-entity layer keys every brand on its root domain and uses the name you saved in Tracked Competitors as the display label across Authority Topics topic breakdowns, the Web Search leaderboard, Citation Gap proofs, the Competitor Leaderboard, and the entity leaderboard. Topic breakdowns also stop mixing in untracked names — only competitors you've actively added show up in per-topic share. Tracked Competitor rows without a domain now show a "Domain needed" chip with an inline input so you can fix them in one click.
Web Search "Competitors tracked" now matches Tracked Competitors — The KPI at the top of Web Search used to count every distinct domain seen in a top-10 result, which made the number balloon (e.g. 113) compared to the 9 competitors you actually opted into tracking. The KPI now reads the tracker directly; the bigger number lives next to it as "Domains observed" so you can still spot new competitors before they show up in your tracker.
Scan status is now accurate at the source level — A scan where every source returned data is marked Completed; a scan where some sources came back empty is marked Partial; a scan where every source failed is marked Failed. Previously the same Completed label could hide a run where all the sources errored out, which let a zero-data run pull the dashboard average down. Scan History also has a dedicated Partial tab so partially-successful runs no longer hide under Completed.
Failed and cancelled scans no longer pollute the dashboard average — The headline visibility score and detection rate on the dashboard ignore scans that finished without data, so a single bad run can't drag the average down. The trend sparkline matches.
Scan results no longer go missing on busy workspaces — On workspaces with several big scans, older scans would sometimes show empty in the detail view even though the data was on disk. We now load each scan's results separately so every scan keeps its full row count regardless of how many other scans are on the page.
Mission Control header now matches the steps below it — When every discovery step is already Synced (you connected sources outside the pipeline or ran them directly), the wave banner now says "Discovery complete" instead of "Ready to start discovery", and the call to action becomes "Generate refined topics" instead of "Run discovery". The same applies to the scoring wave. If the credit estimate ever fails to load, the error block now offers both Retry and "Start anyway" so you're never stuck with no path forward.
Workspace selection now sticks across every page — The active workspace, your plan badge, and the keywords count no longer flash back to defaults when you navigate to BrandCrux, Citations, Socials, Search Console insights, and other dashboards. We persist the workspace in a long-lived cookie so the right one is selected on every fresh load, and every page now reloads its data the moment the workspace resolves.
GEO Audit is now a sixth pillar in your Topic Authority score — Your site's GEO Audit overall score now feeds the unified composite alongside AI mentions, web search, citations, socials, and website coverage. Better schema, semantic HTML, and entity clarity lift your authority on every tracked topic. The BrandCrux dashboard, Topic Map heatmap, and per-topic story pages all surface the new pillar with its own chip and weight. Default weight is 10 of 100 — adjustable per business if you want GEO to count more.
11 May 2026
Keywords now opens sorted by impressions — The Keywords tables land with the highest-demand keywords on top by default. Click any other column header to re-sort.
Bell drawer "View" buttons now land on the right widget — Every notification's View button now opens the exact page (and widget) the alert is about, not the catch-all alerts list. A completed competitor crawl opens the competitor row, a GEO Audit alert opens the report for that audit, a content brief alert opens Content Actions, a social pull opens the matching competitor's posts. Toast clicks follow the same routing.
Search Console insights — easier to scan, clearer when empty — Quick-win queries, declining CTR, intent gaps, CTR vs. benchmark, site sections, and refresh candidates now render as a card grid that wraps cleanly at any width instead of a single dense row. The empty state also distinguishes "Search Console isn't connected for this business" from "connected, no analysis yet" — each comes with the right call to action.
Run AI Visibility scans across multiple personas at once — The New Scan launcher now lets you toggle any combination of personas as chips. Selected prompts run once per persona, with each persona's voice prepended, and the credit estimate updates live with an explicit "× N personas" multiplier so the cost is visible before you approve. Skip all personas to run an aggregate pass with no persona context. Select-all and Clear shortcuts make it one click to fan out or collapse the run.
Real-time progress for every action you approve — The topbar bell + a bottom-right toast now light up the moment a queued action starts running ("Crawling vAuto…"), again when it finishes with the result ("Indexed 47 pages and proposed 3 new topics."), and once more if it fails so you can retry in one click. Notifications arrive within a second of the event instead of waiting for a page refresh.
Content brief suggestions from every GEO audit — When a GEO Audit finishes, BrandCrux now picks the three highest-impact recommendations and offers to turn each one into a full content brief. The suggestions land in the topbar bell drawer with a credit estimate — one click approves, one click dismisses. No briefs ever get written without your say-so, so the audit stays cheap and the follow-up only spends credits when you decide it's worth it.
Dashboards refresh themselves the instant you change something — Add an authority topic, reassign a keyword, or update a persona and the dashboard widgets pick up the change on the next view without waiting for a cache to expire. Previously you could see stale numbers for up to 30 seconds after a change.
Lifecycle Events page for support and engineering — A new page at /app/lifecycle-events shows the audit log of every change that travels through BrandCrux internally — topic edits, keyword reassignments, persona updates, competitor adds, GEO audit completions. Filter chips narrow by event family, and clicking any row expands the full before/after payload. Useful when something should have rippled and didn't.
Discover competitors with AI — proactively — Competitors now has a "Discover competitors" button that asks BrandCrux's discovery engine to surface OTHER companies you should be benchmarking against, not just the ones that already showed up in your scans. It anchors on your authority topics, your products and services, your existing tracked set, and optionally a country, state, or city if your business is local. Each suggestion comes with a one-line rationale, a confidence label, and a one-click Add to tracker.
Scan History — every run, every channel, one page — A new Scan History entry in the sidebar shows every scan you have ever run, across AI Visibility, Web Search, Citations, Socials, Search Console, GA4, and competitor crawls. Filter by channel with the chip strip at the top, sort by started time, duration, status, or credits used, and click any row to drill straight into the full report. Four KPI cards summarise the last 30 days at a glance: total scans, credits used, success rate, and time since last scan.
New Scan is now a multi-channel launcher — Pick any combination of AI Visibility, Web Search, Citations, Socials, Competitor Crawl, Search Console, and GA4 from one page. Each channel has its own parameter section that re-costs live as you tune it, a sticky summary panel shows the total against your balance, and one click queues every selected channel and (optionally) sets a weekly or monthly cadence for the whole bundle.
Competitor social posts now cover every network — LinkedIn, Instagram, X, TikTok, Facebook, and YouTube. Posts from each network land on the Socials page with the right platform icon, and the new Topic Authority scorecard breaks them down by topic × channel × competitor.
Free GEO Audit on every new brand — no setup needed — The moment you add a business, BrandCrux now crawls the first 50 pages of your site breadth-first and scores them for Generative Engine Optimization on us. You'll see a banner on the GEO Audit page while it runs and another the moment it's ready, plus a notification in the topbar bell when the report lands. The free first run never touches your monthly credits.
Personas — score your visibility per buyer archetype — Define the buyer personas you sell to (or let BrandCrux draft them from your business profile). A new persona-scoped scan runs every tracked prompt once per persona, with the persona's voice and pain points framing the question, then tags every result with the persona it came from. The per-persona report shows which prompts surface your brand for each archetype, with per-engine breakdown. Find it under Strategy → Personas; the Persona setup card now appears in Mission Control alongside tracked prompts.
Keywords and Prompts now share one home — They're the same loop (demand → questions), so they live behind one sidebar entry with a tab strip at the top. No work lost; switch between tabs without losing scroll.
Topic dashboards now populate after every scan — A long-standing gap meant some scans completed successfully but the per-topic rollup that drives the BrandCrux composite score and the Authority Topics leaderboard didn't get written. New scans now populate both surfaces the moment each engine finishes.
Search Console keywords auto-map to your topics on pull — When you sync Search Console queries (manually or via the weekly auto-sync), the new keywords now cluster into your authority topics in the same request. No more waiting a week for the topic map to light up.
Every channel now reads as one unit — AI Visibility, Web Search, Citations, Socials, and Owned Analytics each have a sub-tab strip at the top of every page in the channel so you can move between Overview, Competitors, Citations, Posts, Queries, Pages, and Sources without leaving the channel context. New per-channel competitor leaderboards land for AI Visibility, Web Search, Citations, and Socials — same widget bar as the old global Competitors page, scoped to each signal source.
Competitor share of voice now exists per channel — AI Visibility · Competitors shows who appears alongside you in AI answers. Web Search · Competitors shows who outranks you for the keywords your customers search. Citations · Competitors shows who earns directory, review, and editorial coverage you do not. Socials · Competitors shows who posts on your topics and how much engagement they earn.
Socials · Topic Authority scorecard — A new sub-tab in the Socials channel rolls up the last 90 days of competitor posts into a Topic × Channel × Competitor grid. Authority topics lead the table; an "Off-topic" panel surfaces the other subjects competitors are creating content for so you can decide whether to adopt them or use them as content cues. Click any row to see the per-competitor breakdown.
Search Console insights now read like a report — The page now leads with a top-line snapshot (clicks, impressions, CTR, average position vs. the previous 30 days), then surfaces every analysis as its own card: quick-win queries on page two, queries losing CTR, pages serving conflicting intents, keyword cannibalisation, branded vs. non-branded split, plus country and device breakdowns. Each card has its own leaderboard and chips so you can scan the whole report without reading a paragraph.
GA4 insights got the same treatment — Engagement summary as a top KPI strip, channel mix shift and landing-page decline as side-by-side cards, executive summary and recommendations panels carry through. The page now answers "what changed last month?" in one glance.
Authority Cards — live previews of every variant, in one click — The page now shows live previews of all three variants (full card, compact strip, badge) for every card, side by side. Click any preview to open the embed modal pre-set to that style. New help guide explains where each variant fits (landing page, blog post, navbar) and when to pick the JavaScript snippet vs the iframe. The embed snippet box also now reads clean on every theme.
Share an Authority Card as an image — From the embed modal you can now generate a PNG of the selected variant, share it straight to X, LinkedIn, or Facebook, or download it for slide decks. The image is hosted at a permanent URL so the same link keeps working in every channel you paste it into.
Track competitors across every domain they own — A competitor can now have multiple domains (main site + blog + regional TLDs). Open the Competitors page, expand any row, and add as many domains as you like. Every crawl follows all of them so the picture is complete, not just the headline domain.
Stronger competitor crawls when sites push back — If a competitor's site blocks our direct crawl, BrandCrux automatically retries the same domain through a proxy network (capped at 100 pages per domain so costs stay predictable). The crawl history records which domains needed the fallback so the picture is honest.
10 May 2026
The Growth-tier intelligence suite is now reachable from your MCP client — Anyone using BrandCrux through Claude Desktop, Cursor, or any other MCP-aware tool can now ask for the proprietary intelligence layer in plain English. Fifteen new tools cover daily trend snapshots and anomaly hypotheses, classified competitor moves with recommended counters, end-to-end content actions (draft a brief → mark published → measure visibility lift 14 days later), the attribution graph that explains why a metric changed, and embeddable Authority Cards rendered on demand. Same Growth-or-higher gate as the dashboard; agents below Growth get a clear upgrade hint instead of a silent failure.
Trend Intelligence works on day one for new brands — Triggering an on-demand snapshot the first time you opened the page used to fail for newly onboarded brands with no history yet. Fixed; the page now lights up instantly and starts accumulating the time series from snapshot zero.
Mission Control — see your whole brand pipeline at a glance — A new top-level page that visualises the entire brand-authority pipeline as a top-down flowchart of 13 stages across two waves. Wave 1 ingests the evidence (crawl your site, crawl competitors, pull socials, sync owned data). Then we re-recommend authority topics with the new evidence in front of you and surface the diff: which seed topics held up, which weakened, what new topics emerged from competitor content. You confirm refined topics and tracked prompts, then Wave 2 scores everything in parallel — AI visibility, web SERP, directory citations, social authority, owned analytics — and the composite Topic Authority score auto-recomputes.
Embedded onboarding — New users land on Mission Control directly. Tier 0 ("about your brand") renders as a prominent inline form with a grouped category typeahead (50+ standard categories like Auto Dealer, B2B SaaS, Restaurant, plus free-text for anything else). As you fill each prerequisite, the next stage unlocks. No separate onboarding page; the orchestrator IS the onboarding.
One-click bulk pipeline run — A single approval card surfaces the per-thread credit breakdown for the whole wave, with checkboxes to opt out of any specific channel. Approve and Wave 1 fires every discovery thread in parallel; once you've reviewed refined topics + prompts, a second approval card fires Wave 2 the same way. No more manually triggering each scan.
Stage-completion notifications — A new bell drawer in the topbar surfaces in-app alerts for every pipeline milestone (discovery wave done, refined topics ready, scoring wave done, any thread failed). Email alerts fire too — Wave 1 done sends a "review refined topics" email; Wave 2 done sends a "pipeline complete + Topic Authority score" email. Per-user opt-out from settings.
Partial data flows in real-time — Channel pages (Topic Authority, Web Search, Social Authority, Directory Citations, GSC Insights, GA4 Insights) show a banner during active runs surfacing what's complete vs. still in flight. Composite scores recompute as each channel finishes, so you see the picture build up channel by channel rather than waiting for everything.
Resumable end-to-end — Pipeline runs survive any backend interruption. A watchdog cron checks every 5 minutes for stuck stages and re-fires them safely (idempotent — already-running events are no-ops). The state of every wave persists, so a crash mid-flight picks up exactly where it left off without re-paying for completed work or losing partial data.
09 May 2026
Public site restructured around five channels + the closed loop — The home page is now a top-level introduction to BrandCrux as a multi-channel brand-presence platform, leading with The Loop (declare → measure → see → act → re-measure) and the five-pillar grid. Each pillar lives at its own top-level URL: /ai-visibility (the AI Search Visibility deep-dive that used to be the home page), /web-search, /authority-signals, /social-authority, and /owned-analytics. Each pillar page has a "Connected to BrandCrux" section that cross-links to the other channels and back to the composite Topic Authority score.
Pillars dropdown replaces Features dropdown in the marketing nav — The top-level nav now exposes the five channels first, then the AI Visibility sub-features under them. Mobile menu mirrors the same hierarchy. Footer adds a dedicated "Pillars" link section.
BrandCrux — the crux of your brand across channels — The product is now BrandCrux. Same closed loop, broader vision: AI Visibility (the original surface) is now one of five channels under a single brand-presence platform. The other four — Web Search, Authority Signals, Social Authority, and Owned Analytics — sit alongside it, with a composite Topic Authority score per topic that blends every channel into one ranked leaderboard against every competitor we observe. Brand identity (name, domain, email aliases, tagline, social handle) is now sourced from environment variables, so any future rename is a single secret swap.
New landing page: five channels + Topic Authority callout — The home page leads with "The crux of your brand across every channel," shows a five-pillar grid (each pillar links to a dedicated feature page), then a Topic Authority composite band, then deeper coverage of the AI Visibility channel. The FAQ is rewritten to address the broader scope: how this differs from Google Analytics, whether you need to connect every channel before BrandCrux is useful, how Topic Authority blends signals.
Four new feature pages — Web Search (Google / Bing / DuckDuckGo / Yandex rankings with featured snippet, knowledge panel, and PAA capture), Authority Signals (directory + publisher citations across G2, Capterra, Crunchbase, Trustpilot, editorial publishers, forums; Gap List for missed directories), Social Authority (last-90-days competitor posts across Instagram, LinkedIn, X, TikTok, YouTube, Facebook with engagement scoring + content pack generator), and Owned Analytics (Search Console + GA4 deep insights with 7+6 pre-built pivots and 5 specific recommendations per run).
Marketing nav reorganised by pillar — The Features dropdown now groups by pillar: AI Visibility (the existing 8 sub-features), Beyond AI (the 4 new pillars), and Cross-cutting (Integrations). The 9 existing AI Visibility feature pages each get a "Channel 01 · AI Visibility" eyebrow above the hero so they're visually grouped under the new structure.
Favicon, OG image, and meta updated — Browser-tab icon is the BrandCrux teal radar mark in SVG form. og:meta and twitter cards reflect the new positioning. Sitemap + llms.txt rewritten so search engines and AI agents see the elevated product.
May 8, 2026
Sidebar streamlined to four-or-fewer items per group — Navigation rebalanced so no group exceeds four entries: Strategy (Topic Authority, Keywords, Prompts, Competitors) → AI Visibility (Dashboard, Scans, Citations, Trends) → Search Visibility (Web Search, GSC Insights, GA4 Insights, Traffic) → Authority Signals (Directory Citations, Social Authority) → Improve → Setup → Account. Easier to scan, less hunting, and the page title shows once in the topbar instead of being repeated inside the page.
Discover Authority moves into Topic Authority — The discover-from-competitor-crawls wizard is now an inline panel on the Topic Authority page rather than its own separate page. The panel surfaces a count of pending discovered topic suggestions inline, so you can review and accept them without leaving the leaderboard.
Help walkthroughs on every visibility page — Web Search, Directory Citations, GSC Insights, GA4 Insights, and Social Authority each gained a Help launcher (?-icon top right) that opens a step-by-step walkthrough of how the page works, why each signal matters, and what to do with the data.
Sidebar reorganised around the closed loop — Navigation groups now read top-to-bottom as Strategy (Topic Authority, Keywords, Prompts) → Measure (Dashboard, Scans, Citations, Competitors, Trends, Traffic) → Improve (Alerts, Recommendations, Content, GEO Audit) → Setup (Connectors, Settings, Team) → Account (Billing, API, Refer & Earn). Connectors moves out of Workspace and into Setup so the data foundation is more discoverable. Topic Authority leads the Strategy group as the guiding light for everything you measure.
Onboarding now connects Search Console and confirms authority topics before the first scan — Two new optional steps: Demand (connect Google Search Console with one click, paste keywords manually, or skip) and Authority (accept AI-suggested authority topics for your brand). When you reach the prompt-selection step, the suggestions already reflect your real demand signal and the topics you want to own. The first scan runs against a populated, intentional taxonomy instead of a cold one.
May 8, 2026
Visibility expansion — five new dimensions of brand presence, all on one Topic Authority page — Topic Authority now spans every channel that shapes how people find a brand: AI search engines (where it started), competitor website crawl depth, social engagement on competitor handles across Instagram / LinkedIn / X / TikTok / YouTube / Facebook, directory + publisher citations across G2 / Capterra / Yelp / Crunchbase / editorial publishers, and standard web search rank on Google / Bing / DuckDuckGo / Yandex. Each dimension also gets its own dedicated page so a single layer can be drilled into. A composite Authority Score per topic blends all five into one number, with weights you can tune in settings.
Discover Authority Topics from competitor websites — A new Discover Authority wizard lets you pick competitors, set a per-site page budget, and we crawl their websites breadth-first and propose authority topics they're winning that you should consider tracking. Pages are classified into a topic taxonomy in real time; the wizard's last step lists each proposed topic with the supporting URLs and a one-click Accept button that promotes it onto your Topic Authority page.
Directory + publisher citation tracking — A new Directory Citations page surfaces where you and your competitors are listed across directories (G2, Capterra, Yelp, Crunchbase, etc.) and editorial publishers (industry blogs, news, review sites). Per-target citation leaderboard plus a gap list so you can see which directories competitors are listed in that you aren't. New scans run on demand and can be scheduled weekly or monthly.
Standard web search visibility tracker — A new Web Search page tracks where you and competitors rank on Google, Bing, DuckDuckGo, and Yandex for any keyword you choose. Per-keyword position history with sparklines, featured-snippet ownership tracking, knowledge-panel ownership, and a per-competitor SERP-share leaderboard.
Social authority via competitor post analytics — A new Social Authority page connects to your Apify token, pulls the last 90 days of posts from every tracked competitor's social handles across six networks, classifies each caption into your topic taxonomy, and ranks the per-topic engagement leaderboard so you see who's winning the conversation. A built-in caption-pack generator drafts new posts for the brand on any topic, modelled on the highest-performing competitor posts on that topic.
GSC + GA4 deep AI analysis — Two new pages, GSC Insights and GA4 Insights, turn your existing Search Console + Analytics connections into full analytics surfaces. Pre-computed pivots (top decliners, opportunity queries, page-query intent gaps, branded vs non-branded share, cannibalisation, channel mix, top landing-page declines, engagement summary) feed into an AI narrator that produces an executive summary plus 5 specific recommendations on each run. Recommendations land directly in your Recommendations queue.
Pre-approved credit budgets for every scheduled scan — Every scan type (AI scan, competitor crawl, social pull, directory scan, web SERP scan, GSC deep, GA4 deep) now flows through a single pre-approval card: see the credit estimate broken down line-by-line, your balance after the scan, and approve before anything runs. Recurring schedules (weekly or monthly per scan type, configurable per business) auto-pause if your balance dips below one cycle's worth, with email alerts so you can top up before the next run.
Plan credit pools bumped to absorb the new dimensions — Starter is now 6,000 credits/mo (was 3k), Growth is 30,000 credits/mo (was 12k), and Scale is 80,000 credits/mo (was 25k). Free plan stays at 300 credits. Existing subscriptions auto-receive the bigger budget on the next billing cycle. Top-up packs and pricing per credit are unchanged.
May 7, 2026
Mobile-friendly onboarding, full-page scrolling, and a desktop-best banner on shared reports — The 6-step onboarding flow now collapses cleanly on narrow screens: the step strip drops the connector lines, hides labels for steps you're not on, and keeps the active step labelled so you always know where you are. Form fields stack to a single column, the page layout respects the dynamic viewport, and onboarding adds a 400px bottom buffer so the Continue / Back action row is always reachable on any phone. Shared reports (/share/…) now also lead with a clearer top banner on phones explaining that some data tables won't fit a small screen and the report is best viewed on desktop.
Topic Authority — a new page that ranks every topic against every competitor we observe — A new Topic Authority page lets you declare which topics your brand wants to be known for, then ranks you against every competitor we extracted from AI responses. Each topic carries a recommendation badge — Lead, Defend, Compete, Exit — based on your share of voice plus the topic's traffic potential. One-click Generate brief on Compete-flagged topics queues a content brief anchored on that whole cluster.
Recommend authority topics from your data — On Keywords → Manage topics, a new ✨ Suggest authority topics button drafts 8-12 candidate topics based on your business profile, top Search Console keywords, and the competitors we already track for you. Accept the ones you want; the ⭐ marker on each topic flips them into the Topic Authority page automatically.
Page-category and topic badges on every cited URL — The Citations page now classifies each cited URL into a content category (API Docs, Blog, Pricing, Case Study, etc.) and maps it into your topic taxonomy. AI engines lean heavily on docs? You'll see it. Competitors winning on a topic with a single landing page? You'll see that too. Classification runs automatically after every scan and caches per page.
Content-coverage badge on each topic — Each topic header on the Keywords by-topic view now shows how many of your own pages have been cited under that topic. A "No own content" badge flips on for topics where AI engines never reach for your content yet — a clear signal of where to publish next.
Citations page no longer truncates the long tail — Previously the API capped cited pages at 100 across all domains. Own-domain citations on big-volume sites would silently drop off. Now: every own-domain page we ever tracked is surfaced, and external domains get a more generous 250-page cap.
Curated topics — define your own taxonomy and have AI map keywords into it — A new Manage topics action on the Keywords page lets you upload your own list of topic names. When curated topics exist, the AI clustering switches to a constrained-classification mode and routes every keyword into one of your labels (no inventing new ones). Leave the list empty and the workspace continues to free-form cluster as before.
Topic visibility trends — sparkline + 90-day chart per cluster — Each topic header now carries a 90-day visibility sparkline. Expand the topic and you get a chart panel showing how AI visibility for that cluster has moved scan to scan, plus a delta callout: "+12 pts over 8 scans." Built on a per-scan rollup so even a 200-keyword workspace renders instantly.
Brief impact — did this content actually move the needle? — Every brief on the Content page now carries an Impact pill once an artifact ships and at least one scan runs after the publish date. The pill reads "+18 pts on VIN Decoding" (green) or "−6 pts on Pricing" (red) so the loop closes and you can tell whether each piece is earning visibility. Hover for the before/after sample sizes and publish date.
Top competitor pill on every topic — Each topic header surfaces who is winning AI mentions on that cluster (e.g. "Carfax 52%"). Hover to see the top three competitors and their share. Cross-link from your content strategy to the topics where you are losing share fastest.
Topic-level alerts — visibility drops, competitor breaches, content gaps — Three new alert types fire after every scan: Topic visibility drop (cluster fell more than 8 points), Topic competitor breach (a competitor crossed your share for the first time), and Topic position loss (your keywords in the cluster average a position above #8 on Search Console). All land on the existing alerts page.
Keywords now group by topic — and act as a group — The Keywords page has a new By topic view that AI-clusters your positive keywords into semantic groups (e.g. "VIN decoding" or "Vehicle history reports") so you stop scrolling a flat 200-row list. Each topic header carries the cluster's total impressions, weighted AI visibility, and prompt count, plus two one-click actions: Generate brief queues a single content brief that targets every keyword in the cluster as a unit (one strong piece, not 12 thin ones), and Auto-map to prompts proposes which existing tracked prompts each cluster should link to. Suggestions land in a review tray you accept or reject. Topics refresh every Monday automatically and can be recomputed any time with the Recompute topics button.
Generate prompts in your customer's language — The prompt generator now accepts a language picker on the Prompts page. Pick from 20+ languages (Spanish, German, Portuguese, Japanese, Hindi, Arabic, Mandarin and more) and we'll produce buyer-intent prompts in that language so you can measure AI visibility in non-English markets without rewriting them by hand.
Postal codes on Scan Location for tighter local targeting — Add a ZIP / postcode / PIN to your scan and we inject it into the prompt context for both web and API sources. Useful for local-intent prompts where city alone is too coarse — AI engines now anchor recommendations on the actual neighborhood.
Bulk-import competitors from a CSV — On Settings, paste an Import CSV file with one competitor per row (optional second column for domain) and we'll merge them into your competitor list, skipping any that are already there. Good for bringing in a list from a spreadsheet without typing each name.
By Platform headline cards on the dashboard — A new tile strip rolls up web + API visibility into one number per AI platform (ChatGPT, Perplexity, Gemini, Claude, Copilot, Grok, Google AI). The per-engine breakdown still lives below for when you want the detail.
Visibility trend pill on the headline score — The dashboard now shows a percentage change pill next to your visibility score so the direction of travel reads at a glance, not just the absolute delta.
Downloadable visibility report PDF — white-labelled where enabled — The PDF button on the dashboard now produces a real downloadable report (KPIs, by-platform rollup, top prompts, citations, competitors) instead of just opening the print dialog. White-label customers get it rendered with their own brand name, logo URL, and accent color so it's ready to send to clients.
Keyword opportunities now land on your dashboard — The dashboard surfaces the top 5 high-impression search queries where AI engines aren't picking you up yet (or you're not yet tracking them). One click jumps to the Keywords page Opportunities tab to act on them.
Opportunities tab on Keywords — high-impression terms where you're losing — A new Opportunities filter on the Keywords page surfaces positive keywords with at least 100 impressions in the last 90 days where you're either sitting in low search positions, AI engines aren't mentioning you, or you're not yet tracking them. Each opportunity row gets a Generate brief button: one click anchors a content brief on that specific keyword, fed by its real Search Console metrics and the AI visibility we measure for any linked prompt. Bulk-select to queue several at once. The brief lands on the Content page ready to write.
Promote any keyword to a tracked AI scan prompt — one click, then watch the visibility roll in — On the Keywords page, every positive keyword now has an Add as prompt button. Click it and we create a tracked prompt anchored on that keyword, ready to be measured on the next scan. Bulk-select multiple keywords and add them all at once. The keyword row then shows live AI visibility — the weighted average across every linked prompt plus the best and worst scores, so you can tell when a keyword is strong on one engine and absent on another at a glance.
Keyword-to-prompt cross-linking — Expand any prompt on the Prompts page and you'll see chips listing the keyword(s) that drove it. Click a chip to jump straight to that keyword's row in your workspace. Promoted links sit ahead of inferred ones so the manually-curated trail is clear.
Cleaner Keywords and Content surfaces — The duplicated stat-cards-plus-tiny-chips pattern on both pages is gone. Counts now live inline on a single prominent tab strip, with credit-usage and last-brief date demoted to a subtler stat line. Easier to scan, less to hunt for.
May 6, 2026
Downloadable PDF invoices on every successful payment — Click Invoice on any paid row in your billing history to get a professional PDF with full BrandCrux legal details, your billing address, line items, GST breakdown for Indian customers, and the Razorpay payment reference. Replaces the previous plain-text receipt.
Billing details required before checkout — Before you subscribe, top up, or authorize auto-recharge, we now ask for your billing address (one-time, then remembered) and an optional tax ID. The form is country-aware: pick India and we'll validate your GSTIN; pick the EU and we'll ask for your VAT number; the US asks for an EIN; and so on for 12+ countries. Tax IDs land directly on every invoice, so B2B teams can reclaim VAT/GST cleanly without back-and-forth.
Keyword workspace — bring your existing SEO keywords into prompt generation — A new Keywords page lets you add the words your customers actually search. Positive keywords feed customer vocabulary into AI prompt generation; negative keywords are excluded from every generated prompt. Manually paste them, upload a CSV, or click "Suggest from profile" to have AI propose 25 seed keywords for you. On Starter and above, click "Pull from Search Console" to auto-import your top queries from the last 90 days, complete with impressions, clicks, and average position. Newly pulled keywords land in a "Pending review" queue so you accept each one before it influences your prompts. A weekly auto-refresh quietly keeps the workspace in sync with what's working in search.
No more "no data yet" flicker on slow networks — Pages like Scans, Prompts, Competitors, Citations, and Trends used to flash a "run your first scan" empty state for half a second on slow connections before the real data appeared. The app shell now does a single tiny pre-flight call when you switch businesses to learn upfront whether each page has data, so the right view (data, skeleton, or empty state) renders immediately. On a 3G connection that's the difference between a confusing flicker and a clean instant load.
May 5, 2026
Partner Program — earn 25% revenue share for the lifetime of every customer you refer — Marketing agencies, consultants, and creators can now apply to the BrandCrux Partner Program. Approved partners get a unique link, a dedicated portal to track signups and conversions, and earn 25% commission on every paid invoice for the lifetime of the customer. Settlements run monthly; balances above $100 pay out within 7 business days via bank ACH, PayPal, wire, or Razorpay. Partner Program and Refer & Earn run independently — Refer & Earn pays existing users in credits, the Partner Program pays approved partners in cash.
Refer & Earn — get 100 free credits when a friend signs up, 500 more when they upgrade — Every account now has a personal invite link in the new Refer & Earn sidebar entry. Share it however you like. The moment someone signs up via your link, both of you get 100 bonus credits in your top-up balance. If they pick a paid plan, both of you get another 500 credits each. We'll also nudge both sides 72 hours after signup if the referee hasn't upgraded — there's a clear "1,000 credits between you" reminder. Bonus credits never expire. Self-referrals and obvious abuse patterns are blocked.
May 4, 2026
Authority Score now floors to zero for uncited domains — Domains with no own citations were getting an Authority Score in the 30s because the gap and trend sub-scores happily returned high values when there was nothing to compare against. Now if no AI engine has cited you, your score is 0 across both the dashboard widget and the Citations page.
Auto-recharge low-balance alerts fire as soon as you dip — Previously a single alert per 24 hours, even if you crossed the threshold within minutes of a top-up. Now you get notified the first time your balance drops below your configured threshold, then a one-hour cooldown before the next reminder. The 24-hour limit still applies to standard low-credit alerts so they don't pester accounts that just need a manual top-up.
Cleaner upgrade prompts on plan-locked pages — The API Playground and Traffic Analytics pages no longer load a half-empty UI for plans that don't include them. Free and Starter accounts now see a single, honest upgrade card that says exactly which plan unlocks the feature and what it costs. The same prompt-style "unlock with Starter" hint on Web GUI sources in New Scan now shows the right monthly credit count.
Free scan no longer infers the name "Www" when you paste a www.* URL — Submitting www.creditacceptance.com on the free scanner used to produce a report titled "Www" because we'd split the hostname on the first dot. Now we strip the leading www. before inferring, so the report comes back as Creditacceptance.
Signup notifications added to the internal ops feed — Same lightweight admin email system as scans/subscriptions/top-ups, now also fires on every new account creation (email or Google OAuth), with the new account's name, email, signup method, IP, and geo. Toggle with ADMIN_NOTIFY_SIGNUPS.
May 3, 2026
Internal ops notifications — Lightweight admin-only emails now land for every free scan, every subscription event (new signup, upgrade, downgrade, renewal, cancellation, payment failure), and every top-up (manual or auto-recharge). Lets the team spot oddities (a flood of free scans from one IP, a botched checkout, a churned customer) without staring at the dashboard. Configurable per category via the ADMIN_NOTIFY_* env vars.
One scan-complete email instead of two, with the full dashboard report inside — Every scan used to send a separate "alert" email next to the scan-summary email, which felt like double-spam (the alert email's subject even read "Scan completed" — almost identical to the summary's "Scan Complete" subject line). Consolidated to one email that opens with a dashboard-style KPI strip (mentions, detection rate, competitors, engines), keeps the per-engine breakdown table, and now adds two real comparison sections vs your previous scan: Who's Showing Up Instead shows top competitors with delta arrows and a callout when a brand-new competitor appears, and Citation Sources lists the top domains AI engines cited along with how many citations were gained or lost since the last scan. All change alerts that used to fire the second email are now folded into the summary email's Alerts section.
MCP server is now auto-discoverable — Agent platforms (Claude Desktop, Cursor, ChatGPT, anything MCP-aware) can now find BrandCrux automatically through a discovery card at https://api.brandcrux.ai/.well-known/mcp/server-card.json. As more clients adopt the convention you'll see "Add BrandCrux" in their MCP browsers without needing to paste the URL by hand.
Comparison pages — Side-by-side breakdowns of BrandCrux vs Profound, Otterly, Peec AI, Rankscale, and Scrunch are live at /compare. Pricing, feature matrices, sampling-quality wedge, and an honest "when to pick them instead" callout for each.
May 2, 2026
Bigger plans, simpler structure, fresh credits every month — Free goes from 100 to 300 credits (still API sources only). Starter from 300 to 3,000. Growth from 1,200 to 12,000. Scale drops to $179 (was $299) and goes from 5,000 to 25,000 credits. Rollover is gone — every cycle starts with fresh credits, and top-up credits never expire if you want to bank capacity. Top-up packs were also rebuilt: 500 / 1,500 / 3,500 / 6,000 credits at $10 / $25 / $50 / $75. Net for almost everyone: more credits, less to think about.
Enterprise tier is here — Need higher limits, SSO & SAML, audit logs, custom data residency, dedicated success manager with 99.9% SLA, custom integrations, or white-glove onboarding? Reach out to sales@brandcrux.ai.
API + MCP can now manage alerts, billing, and auto-recharge — The public REST API and the MCP server gained a dozen new tools: list and mark alerts read, read your plan and credit balance, list payments and the credit ledger, and inspect or update auto-recharge settings (threshold, amount, currency, on/off, recent attempts, revoke). First-time payment-method authorization still happens in the dashboard because of card-network challenges; everything after that is scriptable. Full reference at /api/docs.
API keys now use the llms_ prefix — New keys are issued with the llms_ prefix to match the product name. Existing keys continue to work as-is.
AI responses now render as clean Markdown — When you expand a prompt result on the Scans page or open a shared scan, the AI's answer now renders from its Markdown form instead of the raw scraped HTML. The HTML version was readable but carried 5–10x more bytes and the source site's styling chrome (utility classes, framework-specific attributes) bled through into the panel. Markdown gives you the same content, in your app's typography, with cleaner headings, lists, and links.
Per-run picker for high-confidence scans — When a prompt is sampled multiple times against the same source, the Scans page and the share report now show a small row of run pills above the response — each pill is one run, color-coded by what the AI said about you (cited & named, cited only, named only, or not found) and tagged with the run's timestamp. Click a pill to load that run's response. Previously the page stacked all 10 runs in a single scrolling block.
Competitors → Ranking Breakdown shows every prompt — The widget was capping the table at 10 rows. It now lists every prompt you track so you can see your position across the full set in one place.
Prompts → expanded view now reconciles top cards with the heatmap — On a high-confidence scan, the source cards above and the historical heatmap below could disagree because each picked an arbitrary sample from a multi-run cell. They now both use the run that actually scored, so the same number appears in both places. The card section also carries a small "most recent scan ([date])" label so you know what time period the cards represent.
May 1, 2026
Time-remaining and notification note on running scans — While a scan is in flight, the running banner now shows about how much time is left (extrapolated from cells already done; says "Estimating remaining time…" until the first cell lands) and a quiet line below it telling you a notification will appear when the scan finishes so you can close the tab. Updates every 5 seconds along with progress.
Cancelled scans now show what they collected — Cancelling a scan mid-flight used to leave the Scans page blank for that run. It now displays every prompt × source result that completed before the cancel, with a Cancelled chip on the row so it's clear the run wasn't a normal failure. The backend also stops enqueueing new model calls within seconds of the cancel and writes the partial output cleanly, instead of letting the run thrash on for several more minutes after you click the button.
Scans no longer double-bill on infrastructure retry — When a scan was interrupted by a transient infrastructure hiccup, the recovery path was re-running every prompt × source cell that had already finished, charging credits a second time and producing duplicate samples. The orchestrator now resumes from where it left off — already-completed cells skip straight through, only the unfinished work runs. No user action needed.
Auto-recharge that actually charges — Auto-recharge is now hands-off. Authorize a payment method once on the Billing page and we'll top your credits back up automatically the moment your balance crosses your threshold. Works with cards (domestic and international) and UPI Autopay; you choose USD or INR; no email-and-click loop, no scans paused mid-run because you missed a low-balance email. If a charge ever fails we email you, retry once a day later, and pause auto-recharge after a second failure so you can update the method on file.
Error monitoring on the way — We now have automatic error capture across the app. When something breaks (a 500 from the API, an unhandled exception in the UI) we get notified in real time so we can fix it without waiting for someone to report it.
Confirm-before-flip on prompt status — The Active / Paused control on a prompt's expanded view is now an explicit "Mark Active" or "Pause prompt" button with a confirmation dialog. The previous single toggle made it ambiguous what tapping it would do; the new flow spells out that activating a prompt adds it to scheduled scans and starts consuming credits, and pausing one removes it from the rotation without deleting data.
Weekly auto-scans on Starter — Starter plans now run weekly auto-scans every Monday. The pricing page reflects it, and the Prompts page has a clear explainer at the top: every prompt marked Active runs in scheduled scans; pause one to keep it out without deleting. Active prompts also carry a small calendar icon next to the query so you can see at a glance which prompts run on the schedule.
Country-aware location field on New Scan — The optional second field next to Country now adapts its label to the selected country: "State" for US / India / Mexico / Brazil / Australia, "Province" for Canada / South Africa / Indonesia, "County" for UK / Ireland, "Prefecture" for Japan, "Canton" for Switzerland, "Region" for France / Italy, etc. Helps users enter the right term for their region without head-scratching.
Watchdog for upstream provider credit — Every 15 minutes we now check the credit balance with our scraping and AI gateway providers. If a provider runs dry, the affected sources are paused immediately so new scans skip them with a clear note rather than burning credits on a guaranteed failure. As soon as credit is replenished, the sources come back online automatically.
Backup scraping path for paid plans — When the primary scraping route can't deliver a cell (rare provider outages, edge cases in their pre-built scrapers), Starter / Growth / Scale scans now have a second path that opens a real browser session against the AI source and reads the response directly. Free plans continue to use the primary route only — the backup is meaningfully more expensive per query, so we keep it gated to paid tiers. Off by default; we'll flip it on per-environment as we validate.
Big scans on the same source go out as one batch — When a scan has 4 or more prompts on the same web GUI source, we now bundle them into one upstream scrape job instead of sending each one separately. Drops the upstream call count proportionally and uses one of the 100 batch slots instead of N of the 5,000 real-time slots, so we have far more headroom under busy hours. Off by default while we validate; flip on per-environment when ready.
One scan per user, faster lane for clicks — Each account is capped at 5 simultaneous scans (across every machine and replica), and user-clicked scans now jump ahead of background work like weekly auto-scans. So a single big customer can't queue up so many scans that everyone else waits, and a click never sits behind a hundred cron jobs.
Web GUI scans now keep their distance from each provider's queue — Each AI source has its own concurrency cap on our side so a heavy scan burst on ChatGPT no longer queues up behind itself and lengthens its own tail. Faster sources (Perplexity, Gemini, Google AI Mode) can still run wide; slower or rate-limited ones (Grok, Copilot, ChatGPT) hold back. Net effect on a normal scan is zero; on big bursts the long tail comes back faster.
Per-source patience is now realistic — Some AI sources legitimately take 7–13 minutes when they're under load, others wrap up in 30 seconds. Every source now has its own deadline tuned to that range (ChatGPT 15 minutes, fast sources 5 minutes) so we recover the long-tail successes without locking up sampler slots on the fast ones.
Quieter ops alerts when the provider is just slow — Our hourly scan-health alert used to email on any partial scan, including ones where the upstream scrape eventually completed but we'd already given up. Now it differentiates: soft outcomes (timeouts and rate limits) are tracked but no longer page support; hard outcomes (captcha, blocked, malformed, content filter) still trigger an alert. Cuts the false-alarm volume materially.
Every scan attempt now carries the upstream scrape ID — When something does go wrong on a web GUI source, the row in our scan-attempts log carries the provider's snapshot ID. Lets us cross-reference a failure with the provider's own logs in seconds instead of guessing.
April 30, 2026
Source cards on the Prompts detail now match the heatmap — On a prompt's expanded view the source cards at the top would say "Missing" for nearly every engine even when the score table below clearly showed scores from the latest scan. The cards were picking an arbitrary older scan instead of the most recent one. They now use the same latest scan the heatmap does, so a prompt that scored 89 on Claude API yesterday shows 89 on the Claude API card today.
Google Search source temporarily off — Google Search has been disabled across scans while we finish the underlying scraping path. Existing scans show their historical Google Search data; new scans will skip it without consuming credits. We'll re-enable it once the new path is live.
Removed Citation Volume Trend from the Citations page — The trendline at the top of the page wasn't adding signal beyond what the Citations Over Time and Citation Sources widgets already cover. The page is now tighter.
Slow ChatGPT and Gemini scans no longer get marked failed — Web GUI scans (ChatGPT, Copilot, Perplexity, Grok, Gemini, Google AI Mode) now use an async trigger + poll workflow against our scraping provider. A single long job (we've seen ChatGPT runs go 5-13 minutes when the model is slow) used to time out our HTTP connection at 3 minutes and report a failure on our end even though the scrape was finishing successfully on the provider side. The connection no longer stays held open for the full job, per-call timeouts are tight (transient blips no longer abort the whole run), and the overall budget is 10 minutes — so the long tail of slow runs now lands as completed cells with full data instead of "scraping failed".
Admin can actually grant credits and change plans now — Plan switches and credit grants from the admin Users page used to fail silently because the UI tried to update other people's rows directly through a path that's blocked for safety. They now route through a privileged server endpoint that updates both the legacy plan column and the canonical subscription plan column, sets the right monthly credit base, business cap, and resets the cycle counter. Quotas now match the published plans (Free 100 / Starter 300 / Growth 1,200 / Scale 5,000), Starter is in the dropdown, and every change is written to the audit log.
Grant up to 100,000 credits in one shot — The admin credit dialog now goes up to 100,000 per action with one-click presets at 100, 500, 1,000, 5,000, 25,000, and 100,000. Negative numbers revoke. Each action shows the user's current balance and the new balance after the change.
Optional email when an admin changes a plan or credits — Both admin actions now have a "Email this user" checkbox with an optional note. If you tick it, the user gets a clean email from billing@brandcrux.ai summarizing the change and your note. Off by default — silent updates stay silent.
Refine drafts with one credit — The article editor has a new Refine panel. Type what you want changed ("rewrite the comparison table with verified 2026 data", "tighten the intro", "replace the section on X") and the article is re-written by the same model that produced it, then reviewed by the same critic if one was used. Costs 1 credit per refinement and saves a new version every time.
Drafts editor now actually shows the draft — Opening a generated article (Markdown body, Meta Title, Meta Description, slug, focus keyword) was loading blank because the editor was looking at the wrong field. The split editor, preview pane, SEO panel, and version selector now populate correctly. Save also writes back to the right fields.
Cleaner slugs and full meta titles — Article slugs are now proper URL slugs (no colons, question marks, or stray punctuation; collapsed hyphens; lowercase). Meta titles are no longer hard-truncated mid-word at 60 characters; you see the full title in the editor with a live counter showing when it exceeds the SEO recommendation, so you can trim it the way you want.
Generate Image and Export now work in the editor — The Generate Image button on a draft now correctly fires the image generation flow with a prompt derived from the article title. The Export button no longer silently closes its own menu the moment you click it; Markdown, HTML, and JSON export all work.
Removed the Competitor Mentions Over Time chart — The trend lines on the Competitors page often showed a single dot or noisy data and were not adding signal beyond the leaderboard. The page now goes straight from Share of Voice into the per-competitor leaderboard with threat scores and head-to-head breakdowns.
Per-sample dots and selectable Confidence on every scan — Each prompt × source row in the scan detail now shows tiny green/grey dots (one per sample run) so you can see attempt outcomes at a glance. The scan setup form has a new Confidence selector (Standard / High / Maximum = 1× / 2× / 3× samples per cell). The credit estimate updates live with the multiplier so you always see the exact final cost before you queue. Higher tiers tighten the Wilson confidence interval (Standard ±15-25% → Maximum ±7-10%).
Scan detail now shows every prompt for every engine, even failed cells — Previously, prompts where every sample retry failed silently disappeared from the per-engine table. So an engine with a high failure rate looked like it had only run 2 of 15 prompts. The view now fills in the missing cells with a "Failed — all retries failed" row, so the per-engine column is always the same length and operational issues are visible instead of hidden.
Web GUI scraping failures now retry properly — When the browser-capture path fails for a query, the error string ("All scraping tiers failed for X") used to be classified as "unknown" and skipped — no retry. Now it's classified as an upstream error, which is in the retryable list, so transient hiccups don't drop the cell on the first attempt.
"What Changed" banner is gone when there's nothing to say — Previously a scan could land an empty banner with just the business name ("Marketcheck Cars Inc.'s visibility") because the narrative generator returned a fragment. The dashboard now hides the banner unless the narrative is a substantive sentence; the email digest does the same; the narrative generator can also explicitly return NONE when there's nothing meaningful to report.
Scan detail no longer repeats the same prompt 10 times — Multi-sample scanning (we run up to 10 attempts per query × engine cell to compute a confidence interval) was leaking through the UI as one row per sample. The expanded scan view now aggregates samples into a single row per prompt, with a small "found / total" sample badge (e.g. "1/10") next to the status when more than one sample ran.
Prompts page splits Active and Paused — A new Active / Paused / All filter on the Prompts page makes it clear which prompts are part of ongoing scans (Active) versus generated-but-not-selected or manually disabled (Paused). Defaults to Active so the table opens on the prompts that actually run.
Competitors page now aggregates across the last 5 scans — Latest-scan-only under-reported competitors for businesses with mature scan history (e.g. Marketcheck Cars Inc dropped from 49 detected competitors to 9). Threat scores, win/loss counts, and the per-competitor leaderboard now use up to 5 recent scans for richer signal.
Competitors and All Entities are separate tabs again — Competitors is the default tab (KPIs, per-competitor leaderboard, head-to-head, threat scores). All Entities is the secondary tab (every brand AI engines mentioned). Onboarding-listed names get * in both.
Restored the original Competitors KPIs and per-competitor leaderboard — Brought back the "Competitors Found / You Win / You Lose / Both Beat" KPIs and the per-competitor leaderboard with threat scores and per-query head-to-head breakdown that were briefly removed when the All Entities view became the default. Both views now render together on the Competitors page: the per-competitor leaderboard at the top, then the AI-detected entity leaderboard below. Onboarding-listed competitors are flagged with * in both.
Phantom self-mention bug fixed on the Competitors page — A scan would sometimes show "Your Share of Voice" as 1% on ChatGPT (or any engine) even when your brand wasn't mentioned at all. The entity extractor was emitting a zero-mention placeholder for the tracked brand which leaked into the share-of-voice math. Both the dashboard widget and the analytics endpoint now skip these placeholders.
Headline percentages no longer trail with 5.19999999… — The "Your Share" KPI and the per-entity share percentages on the Competitors page are now rounded to 2 decimal places.
Optional location injection for Web GUI prompts — A new checkbox in onboarding and the scan setup form lets you prepend "I'm based in [city, state, country]." to the query that gets typed into ChatGPT Web, Perplexity Web, Gemini Web, etc. Off by default — Web sources normally use proxy IP geolocation, which is enough for country-level. Turn on if you need stricter region grounding (e.g. you want India-specific recommendations even though the model might default to US-leaning answers).
Partial scans now show up everywhere — Dashboard, Prompts, Competitors, Traffic, Trends, Citations, Scans, and shared report pages now treat a "partial" scan (one where some engines didn't finish) as a regular completed scan for display purposes. You no longer see "No scan results yet" right after a scan that successfully ran 19 of 20 cells.
AI Assistant icon visible to every plan — Free and Starter users now see the sparkle chat icon in the top bar. Opening it shows a locked card with example questions ("What's my visibility score on ChatGPT this week?", "Find the top 5 citation gaps to close", etc.) and a clear path to Growth. Growth and Scale users keep the full chat exactly as before.
Hourly scan health check and daily digest emails — A new background check runs at the top of every hour, scans the last hour for any "partial" scans, and emails support@brandcrux.ai a per-engine OK/Failed/Fail-% table only when something actually failed (no daily noise on healthy hours). A separate daily digest goes to anand@brandcrux.ai with totals, status breakdown, and per-engine success rates over the previous 24 hours.
API-tier scans now retry every cell up to 3 times — Previously, API engines retried twice and Web GUI engines retried three. Bumped the API tier to match — same cap on both — so transient gateway timeouts have one more attempt before a cell is dropped.
Scan setup form polish — The City input no longer overflows the card on narrow viewports, the "Run Scan" button no longer silently does nothing for paid users with more than 5 prompts (the 5-prompt cap was a Free-tier guard that leaked into the paid path), and Website is now required for both online and local businesses.
Competitors page is now one unified leaderboard — Dropped the "Competitors / All Entities" tab split. Every brand AI sources actually mentioned for your prompts shows up by default. Brands you also listed as competitors during onboarding are marked with a *; a footnote explains the asterisk and clarifies that the onboarding list is a reference, not a filter.
Competitors page now opens on All Entities — The default tab now shows every brand AI engines mentioned for your prompts, not just the small list you typed in during onboarding. So all the local competitors (e.g. Groww, Zerodha, Scripbox for an Indian market scan) show up by default — no toggling required.
"Citations by Source" widget tells the truth when you're uncited — When no engine cites your domain yet, the subtitle now reads "Engine citation activity (none cite your domain yet)" instead of the misleading "Which AI engines cite you most".
Authority Score Trend stays neutral when there's nothing to trend on — A new business with two zero-citation scans is no longer flagged as "declining"; we report 50 (neutral) until there's actual citation history to compare.
Citation Rank now reads zero when you aren't cited — Previously, an uncited domain produced a perfect Citation Rank of 100 (a math edge case where rank=0 collapsed to the top of the percentile). Authority Score now correctly reflects an uncited state across all five components.
Claude API scans no longer drop slow cells — Slow Claude responses (sometimes 100–180s on long, source-heavy prompts) used to fail with an unrecognized error string and weren't retried. We now classify these correctly as timeouts (so they retry up to 3 times) and gave Claude a 180-second per-attempt budget so most slow cells succeed on the first try.
City and state targeting for API sources — Scans now accept City and State alongside Country. We expand country codes to full names ("IN" → "India") before sending them to AI sources, so you no longer get US-flavored results when targeting India. City/state targeting is most accurate on API sources; Web GUI sources only attempt country-level targeting.
Scan summary emails now show competitors and next steps — The post-scan email surfaces the competitors AI sources mentioned for your prompts ("Who's Showing Up Instead") and links you straight to the three actions that move the score: review recommendations, find content gaps, and run a GEO audit. All of it can be done inside BrandCrux.
Cleaner engine names in scan emails — Engine rows now read "ChatGPT", "Claude", "AI Overview", and "Google Search" instead of raw API IDs like "Openai".
Web GUI sources now require Starter or higher — Free plans run scans against the 8 API sources (ChatGPT, Claude, Gemini, Perplexity, Grok, DeepSeek, Mistral, Copilot APIs). Web GUI sources (ChatGPT Web, Perplexity Web, Gemini Web, Copilot Web, Grok Web, AI Overview, Google Search) unlock with Starter ($29/mo). Onboarding and scan setup show locked Web sources with a clear upgrade path so you always know what your plan includes.
Smarter onboarding scan setup — The first scan now defaults to API sources (1 credit each) for Free plans so your first scan always fits in the 100-credit allowance. You can toggle individual sources on or off and see live credit math before you run.
Fixed onboarding showing both "Scan failed" and "Scan running" — When a scan failed during onboarding, the page sometimes rendered both the error and the queued-success state side by side. The error state now hides the queued message so the result is unambiguous.
Removed Google Maps as a tracked source — Google Maps is no longer offered as an AI source across onboarding, scan setup, dashboards, share reports, integrations, and pricing pages. We focused our coverage on the AI assistants and SERP surfaces where mentions actually move recommendations.
April 29, 2026
Publish to 5 CMS platforms from Content Hub — You can now push AI-generated content directly to WordPress, Shopify, Webflow, Wix, and Squarespace. Connect your CMS from the Connectors page, then use the new "Publish to CMS" dropdown on any content artifact. Each platform uses API token authentication with no OAuth flow required.
CMS integration guides — Step-by-step setup guides for each CMS connector are now available at /guides. Each guide covers prerequisites, credential generation, connecting in BrandCrux, and troubleshooting. A link to the relevant guide also appears inside each CMS connect modal.
Team members now inherit the business owner's plan features — Previously, invited team members were blocked from Growth/Scale features like Content Hub, Connectors, and AI Assistant because the system checked their own plan instead of the business owner's. Team members now see and use all features available on the business they are accessing.
April 29, 2026
API and MCP coverage for the new sampling fields — The public REST API and the MCP server now expose every confidence-aware field added this session. GET /api/v1/scans/{id} returns presence_rate, Wilson CI bounds, sample counts, quality_flag, mention_rank, score stddev, and per-sample scores per cell. New GET /api/v1/scans/{id}/attempts surfaces per-pull attempt diagnostics. GET /api/v1/visibility-score includes a confidence block with degraded/suspect/insufficient cell counts. New GET /api/v1/source-recommendations and POST /api/v1/competitors endpoints. MCP gains get_scan_attempts, add_competitor, get_source_recommendations, and get_data_quality_summary tools so AI assistants can reason about measurement confidence.
Confidence-aware Ranking Breakdown — The Dashboard and Competitors page rank widgets now show per-cell sample counts (e.g. "5/5"), data-quality dots (degraded / suspect / insufficient), and a 95% confidence interval on the aggregate "Avg Rank when listed" KPI. Insufficient cells are excluded from averages. The InfoTip now clarifies that rank means "position in this answer's ordered list," not a fabricated global AI rank.
Methodology article and landing-page explainer — A new "How BrandCrux Measures AI Visibility" article walks through multi-sample probing, Wilson confidence intervals, decomposed judging, and quality flags. Linked from the landing page and the AI Visibility feature page so you can see exactly how our numbers are produced.
Citation Gap Analysis is now the canonical "earn coverage" view — The aggregated Source Recommendations card has been removed from the Citations page. The Citation Gap Analysis section already surfaced the same signal in a more useful form: for each (domain, query, engine) where AI cited a third-party domain and named your competitors but not you, you see exactly which publisher to chase and why. The programmatic source-recommendations endpoint remains available for API consumers.
Confidence-aware visibility scoring — Every visibility cell is now sampled multiple times per scan and rolled up into a presence rate with a 95% confidence interval. Higher tiers get more samples for tighter bands. The dashboard surfaces "based on N attempts" so you know how solid each number is.
Adaptive deeper sampling on uncertain queries — When a query is on the edge (mentioned roughly half the time), we automatically run more samples to lock in a tighter result. Queries that are clearly always-mentioned or never-mentioned stay efficient.
Data quality flags — Every cell is tagged ok / degraded / suspect / insufficient based on how many attempts succeeded and what failed. Suspect cells (where captchas or refusals dominated) are flagged rather than silently averaged in.
Source Recommendations — A new section on the Citations page ranks the third-party domains AI sources cite most in your category, with multi-engine consensus and breadth signals. Each row links to a quick coverage check so you can see if you appear there yet.
Two-stage judge for sharper classification — The visibility judge is now split into a structural pass (entities + mention type) and a sentiment pass that only fires when something was mentioned. More reliable on edge cases like "neutral but listed" vs "actively recommended".
Versioned extractions — Each scan result now stores the verbatim engine output alongside a judge fingerprint. When we improve the judge, we can replay it on history and compare apples to apples.
Per-attempt diagnostics — Every fetch attempt — successful or not — is logged with a categorised outcome (timeout, captcha, content filter, network error, etc.). When an engine has a bad day, we can tell you exactly what failed instead of papering over it.
Query topic and intent grouping — Queries can now be tagged with a topic (e.g. "CRM software") and an intent (informational / comparison / recommendation / problem-led / transactional / local), with paraphrases linked back to a canonical query. Reporting rolls up at the intent level for stable signals.
April 29, 2026
AI Assistant actions — The AI Assistant can now take actions on your behalf: run scans, add/pause/delete tracked prompts, generate content briefs, create content, and update business details. It always asks for confirmation before executing. All actions are security-scoped to your account only.
Shared Share of Voice component — The Dashboard, Competitors page, and shared reports now all pull from a single backend endpoint, eliminating ranking mismatches between pages.
Smarter traffic comparisons — Period-over-period deltas on the Traffic page are suppressed when the comparison period has insufficient data (e.g. newly connected connectors), preventing misleading +74,000% changes. Traffic page now defaults to 14 days.
Send Feedback — Added a "Send Feedback" option to the user menu in the sidebar for quick access.
Upgrade modal light mode fix — The upgrade prompt is now properly themed in light mode with readable text and visible buttons.
AI Assistant side panel — Open the AI Assistant from any page using the floating chat button in the bottom-right corner. Ask questions about your visibility data without leaving your current workflow. The assistant knows which page you are viewing and tailors its responses accordingly. Available on Growth and Scale plans.
Login page layout fix — The sign-in page no longer appears squished on narrow browser windows or when accessed via invite links. Google sign-in now correctly redirects to the invite page after authentication.
Fortnightly GEO audits — GEO audits now run automatically every other Wednesday night. You'll receive an email Thursday morning with your updated scores, category breakdowns, and changes since the last audit.
Weekly auto-briefs — Every Tuesday, BrandCrux automatically generates fresh content briefs based on your latest scans and newly discovered visibility gaps. You'll get an email when new briefs are ready. Available on Growth and Scale plans.
Content Hub guide and focus areas — The Content Hub now has a detailed info tooltip explaining the full workflow. Creating a brief lets you pick a focus area (Product, Content Gap, Topic Authority, FAQ, Comparison, Case Study) which pre-fills the content type and priority.
Model picker for content generation — Choose which AI model generates your content. Pick from free models like Gemini Flash for quick drafts, standard models like Claude Sonnet and GPT-4o for polished content, or premium models for deep research and critique. Each shows its credit cost upfront.
Current-year content recommendations — Content advisory and generated briefs now always reference the current year and date. No more "2024" in titles when it's 2026.
Traffic by property breakdown — When multiple Google Analytics or Search Console properties are connected, the Traffic page now shows a per-property comparison table at the bottom. Compare sessions, clicks, organic %, and more across different properties side by side.
Detailed data tooltips — Every data widget on the Dashboard and Traffic pages now has a rich info tooltip explaining what the data means and how each column is calculated. Hover the (i) icon on any card title to learn more.
AI Assistant — Ask questions about your visibility data in natural language. The assistant looks up your scores, trends, competitors, citations, and content briefs in real time. Powered by Claude Sonnet. Available on Growth and Scale plans (1 credit per 10 messages).
Shareable conversations — Share any AI Assistant conversation as a read-only link. Recipients can view the full thread without logging in.
MCP connection guide — The assistant page now promotes connecting BrandCrux to Claude Desktop, ChatGPT, or Cursor via the MCP server for direct data access from your favorite AI tools.
Inline visualizations in AI Assistant — The assistant now renders score badges, bar charts, line charts, pie charts, and data tables directly in chat responses for at-a-glance insights.
Dark/light theme toggle on website — All public pages now support dark and light themes with a toggle in the navigation bar. Defaults to your system preference.
Mobile responsive overhaul — Every app page now adapts to mobile screens. Compact topbar, collapsible grids, scrollable tables, and touch-friendly controls across all 22 pages.
GEO Audit default view — The GEO Audit page now opens with your most recent audit results displayed by default, with audit history shown below.
AI Assistant feature page — New dedicated landing page at /features/ai-assistant showcasing example queries, visualization types, MCP integration, and credit pricing.
April 28, 2026
AI traffic tracking — See exactly how many visitors ChatGPT, Perplexity, Claude, Gemini, and other AI engines send to your website. Per-engine breakdown, per-page heatmap, and week-over-week trends.
Google Analytics & Search Console connectors — Connect GA4 and GSC with a property picker. 90-day historical backfill on first connect, daily auto-sync, and email notification when data is ready.
AI visibility vs traffic correlation — Side-by-side view of your AI visibility score and actual referral traffic per engine. See which engines recommend you AND send visitors.
Content advisory enrichment — Recommendations now factor in AI traffic patterns. "This page gets ChatGPT traffic but has high bounce rate" type insights.
Cross-referencing insights engine — Automated insights comparing AI visibility with organic traffic. Detects opportunities, validates strategies, and flags drops.
April 27, 2026
Design system overhaul — Completely new UI built on oklch color tokens. Refined sidebar with active indicator bar, breadcrumbs in topbar, density modes, and consistent component styles across all 22 pages.
Executive Report page — New CXO-ready report with revenue opportunity calculator, competitive threat matrix, topic authority analysis, ranking breakdown, and 3 Quick Wins.
Shared report enhancements — Added topic authority, ranking breakdown per source, AI traffic section with blurred mockup CTA, and "more recommendations" prompt. Mobile-friendly layout with logical section ordering.
Scans page rewrite — Scan history view with expandable detail, cancel/re-run/delete actions, animated progress for running scans, and auto-reload on navigation.
Prompts heatmap — Score history shown as a color-coded 2D grid (sources × dates) with inline sparklines per engine. Replaces the old flat chip list.
Scan queued modal — Beautiful overlay modal when a scan is queued, with "View scan status" action.
Business context sync — Selected business persists across page navigation via sessionStorage. Switch transition overlay shows "Switching to [Business]..." during business change.
Entity deduplication — "Marketcheck", "Marketcheck Cars Inc", and "MarketCheck" now merge into one entity with combined mentions.
Smooth trend lines — All trend charts now use catmull-rom bezier curves instead of straight polylines. Hover tooltips on data points show date, engine, and value.
April 26, 2026
Date range selector — Dashboard now has a preset dropdown (7d, 14d, 28d, 30d, 90d, this month, last month, quarter to date) plus custom date inputs. Selection persists across pages. Full page skeleton reload on change.
PDF export — Export any report page as PDF via print. Added to dashboard, prompts, trends, competitors, entities, citations, scans, and executive report.
Toast notifications — Real-time toast popups in bottom-right corner for new alerts. Polls every 30 seconds, auto-dismiss after 8 seconds, click to navigate.
Delete business & account — Delete a business (with all data cascade) or your entire account from Settings. Name-confirmation required for safety.
Team invitation improvements — Resend pending invitations, non-BrandCrux users see Sign In/Sign Up on invite page, auth redirect preserves invite token.
API documentation portal — Comprehensive docs with sidebar navigation, 5 use cases with code examples, "Copy for LLM" button, per-endpoint copy, rate limits, and error codes reference.
Connect CTA prompts — Pages that benefit from GA4/GSC data show dismissible banners prompting users to connect Google.
April 25, 2026
Team collaboration — Invite team members with roles (Owner, Admin, Member, Viewer). Share businesses across your team.
Onboarding email sequence — New users receive a 4-step guide: welcome, score explanation, competitor insights, and usage tips.
API visibility blog — New article: "Why Tracking Your Visibility in LLM APIs Is More Important Than Web Chat"
April 25, 2026
Auto-discover new AI models — BrandCrux now automatically discovers and adds new language models from 10 major providers daily. Each model includes pricing, capability tags, and a "why track this" explanation.
Provider pages — Dedicated pages for each AI provider (OpenAI, Anthropic, Google, etc.) showing all tracked models and integrations.
Integrations page — Browse all 45+ integrations across AI sources, analytics, content publishing, notifications, and developer tools.
Why API Visibility section — New landing page section explaining why API-channel tracking matters for the agent economy.
April 24, 2026
BrandCrux rebrand — Formerly Prism. New name, new domain (brandcrux.ai), new icon.
Scan completion emails — Get a summary email after every scan with per-engine breakdown and comparison to previous scan.
GEO audit email notifications — Receive detailed audit results by email with category scores and page-level issues.
Scan frequency control — Choose daily or weekly automated scans per business.
Light mode overhaul — Improved contrast, depth, and visual clarity for light theme users.
April 24, 2026
AI scorer improvements — More accurate implicit mention detection with confidence thresholds, reducing false positives.
Free scan query selection — Choose which prompts to scan instead of auto-selected queries. Add your own custom prompts.
Device-based rate limiting — Fair usage controls for free scans and signups.
Browser error logging — Development tool for faster bug detection.
Earlier
18 AI source support — 9 Web GUI + 9 API engines tracked simultaneously.
Competitor tracking and share of voice — See which brands AI recommends instead of you.
Citation and entity analysis — Track which domains are cited and which entities are mentioned.
GEO Audit for site optimization — Crawl your site and score every page for AI citation readiness.
Content Engine — AI-powered briefs, writing assistance, and publishing workflows.
Google Analytics & Search Console integration — Connect GA4 and GSC for holistic visibility data.
WordPress publishing — Publish optimized content directly to WordPress from BrandCrux.
REST API and MCP Server — Programmatic access to all scan data and platform features.
Weekly digest emails — Automated weekly summary of visibility changes and competitor movements.
Credit-based pricing with auto-recharge — Flexible pay-as-you-go credits that never expire, with optional auto-recharge.