Changelog

The Insights "know" cards — Recall, Top-of-mind, and Fact accuracy — now chart every run you've made: a sparkline, the latest score, and how much it moved. Run them monthly and watch whether your work is actually sticking in the models' memory.

The sidebar now separates Rank tracking (your domains) from AI visibility with clear group labels — no more hunting for a site in the middle of AI links. Keyword history also got cleaner: failed checks now show a simple "credits refunded" note instead of technical internals.

Track whether Bing's Copilot AI answers recommend your brand — the ninth engine, included in Pro and Agency plans at the low SERP-surface rate. Like Google AI Overviews, an empty result is a signal too: Bing chose not to show an AI answer for that prompt.

Gemini checks now run on Google's latest generally-available model — Google retired the previous one on their side, which briefly made Gemini runs fail (always auto-refunded).

You can now move a domain between groups from its settings — keywords and history ride along. Fixed doubled headers on two tool pages. URL tools now explain nicely when an address can't be checked instead of showing an error page, page reading got more reliable, and refunded AI Lab runs are rarer thanks to sturdier answer parsing.

The Insights page now tells the story in three acts: what engines KNOW about you (your latest Recall, Top-of-mind, and Fact-accuracy results now persist and show here, with age and one-click re-runs), what they FETCH about you, and how STABLE it all is. Three new free cards too: Mention prominence (lead recommendation or footnote?), Co-citation neighbours (who answers lean on next to you — and who owns the room when you're absent), and your own answer-volatility chart per engine.

Prompts in Slovenian, German, and 30+ other languages now carry an explicit answer-language instruction to every engine. Short prompts were ambiguous between related languages — a Slovenian prompt could come back in Czech. Cached answers refresh within a day.

Prompt Fan-out maps the sub-queries an AI actually runs behind your question and flags which facets your tracked prompts don't cover yet — with one-click adding. Top-of-mind Check asks every engine, from bare memory, which brands it names first for your category. Plus two free tools: a public AI Answer Volatility index (how much each engine's answers changed day over day, from our aggregate tracking data) and an AI Facts File generator — llms.txt's sibling for canonical brand facts, with ready-to-paste JSON-LD.

Two new Insights cards: every AI answer classified into the four mention×citation states — including the Evidence Gap, where your content fed the answer but someone else got the credit — and a per-engine breakdown of WHAT each engine cites (forums, reference sites, news, reviews…), so you invest where that engine actually looks.

Brand Fact Audit (what each engine believes about you from memory — and where it's wrong), Recall Check (memory vs. search: is your visibility durable or rented?), Selection Duel (which of two pages an AI would pick as its source, judged in both orders to cancel position bias), and Snippet Lens (exactly which sentences an AI would lift from a page). Credit-priced, auto-refunded when nothing usable comes back.

Type your brand and one question on our homepage and see the actual AI answer — with your brand highlighted, or conspicuously missing while a competitor gets recommended. Two free checks a day.

Every alert digest is now permanently recorded per project — open a project's Alerts page to see everything that ever fired, long after the bell was cleared.

Track whether DeepSeek and Mistral's Le Chat recommend your brand — both included in Pro and Agency plans at the low engine rate, no add-on fees. Prompt generator also went worldwide: one clean English template set plus an optional market field.

The prompt generator now speaks four languages (EN/SL/DE/HR) and can localize prompts to your market; the CTR estimator got a more conservative curve; the Tools sidebar section collapses like the project tree.

Research ANY brand's AI visibility — even ones you don't track. Search a large corpus of real AI prompts and answers for mentions, sample answers, and the sources powering them. Credits per lookup, auto-refunded when nothing comes back.

"Can AI read this page?" (JS-shell detector), an AI prompt library generator, a CTR estimator with an AI-Overview toggle, and a maintained AI crawler list with copy-paste robots.txt rules.

Page titles now condense out of a subtle violet mist. Blink and you'll miss it — that's the point.

PDF reports got a redesign: executive summary, a visibility trend chart, movement arrows on every keyword, and citation-share bars — the same data, dressed for a client meeting.

The project tree no longer folds shut every time you click into AI visibility, Insights, or a domain.

Tag prompts by persona, product line, or funnel stage — then filter the whole AI visibility dashboard, every chart and score included, to just that segment.

Eleven guides covering everything from your first rank check to reading the perception map — public, searchable, and linked from the app sidebar.

Add your company name, address, and VAT ID on the Billing page — every invoice carries them, valid EU VAT IDs bill with reverse charge, and PDFs download straight from the new Invoices section.

The credit ledger now speaks human: "Monthly plan credits", "AI visibility check · run #5" — instead of internal codes.

Scheduled Google AI Overview checks are processed in the background queue — same data, lower cost, which keeps credit prices where they are.

Insights now shows exactly which of your URLs the AI engines cite, plus Citation Gaps: source pages where competitors are mentioned and you are not.

AI Bot Access Checker (does your robots.txt block the crawlers behind ChatGPT, Claude, Perplexity?) and the AI Citability Analyzer (paste content, get a citability score with fixes).

Every project can generate a read-only rankings page for clients — no login needed, revocable anytime.

Purchase confirmations after checkout, and the trial banner now names exactly which module is still on trial.

Every domain now gets a monthly snapshot of ALL keywords it ranks for — across every country — included free with every plan. One-click tracking for anything worth watching.

Change a keyword's text, country, or language without deleting and re-adding. History stays.

The new bell shows run alerts even if you haven't configured email or Slack — including the new "Rising rival" alert when an untracked brand keeps appearing in AI answers.

The Insights page: AI Visibility Index, brand perception map, rank-vs-AI quadrant, and concrete recommended actions.

AI-readiness analyzer, llms.txt generator, and the GA4 AI-traffic setup guide — free for everyone.

Agency plans can brand reports with their own logo, colors, and client name.

Rank tracking and AI visibility in one product: keywords across Google, brand presence across ChatGPT, Gemini, Claude, Perplexity, Grok, and Google AI Overviews.