# Attensira > Attensira is software that does your AI-search visibility work. A crew of AI agents learns your business, picks the questions your buyers ask ChatGPT, Perplexity, Claude, Gemini and Google AI Overviews, and does the work to get you named: research, writing, page fixes, placements and ads, shipped to your site as pull requests or CMS drafts. It is built for founders and marketing leads at SaaS companies, agencies running several brands, and professional-services firms. Nothing ships without your approval, claims always wait for you, and every rate it reports carries the sample size it was measured from. > > Whatever it takes to win the answer. > For the complete documentation index, see [llms.txt](https://attensira.com/llms.txt). > Every page on this site is also served as Markdown: append `.md` to any URL, or send `Accept: text/markdown`. **What Attensira does.** Attensira is software that does your AI-search visibility work. It reads what AI assistants answer about a brand, works out why the brand is missing from the answers it wants to win, and then a crew of AI agents does the work: research, articles and page fixes in the brand's voice, pull requests or CMS drafts, pitches to the pages the answers cite, community replies drafted for a person to post, and Google Ads campaigns within a set budget. Workspaces start in Draft, so nothing leaves the workspace until a person approves it, and anything that states a claim waits for approval in every mode. It reads answers from 5 AI engines: ChatGPT, Perplexity, Claude, Gemini and Google AI Overviews. It does not track any other platform. **Vocabulary.** These are the exact nouns Attensira uses in its app, REST API and MCP server. They match the terms used across the AI-visibility category, so a tool call written against another platform's vocabulary will usually be right here too. - **Project** — one tracked brand and its domain. An API key is scoped to one project. - **Prompt** — a question tracked on the AI platforms, e.g. "best ETL tool for startups". - **Topic** — a named bucket of prompts, reported as one thing. A prompt has at most one topic. - **Platform** — an AI product an answer is read from: ChatGPT, Perplexity, Google AI Overviews, Claude, Gemini. - **Answer** — one AI response to one prompt on one platform. - **Mention** — the brand named in an answer's text, with or without a link. Defined at https://attensira.com/glossary/ai-brand-mentions. - **Citation** — a linked source in an answer. A citation is not the same as a mention. Defined at https://attensira.com/glossary/source-citation. - **Source** — a domain cited in answers, ranked by how often it is cited. - **Competitor** — a rival brand tracked alongside yours, reported as its own mention rate. - **Visibility** — a label on a section of the product and a value in the `get_analytics` area enum. It is not a computed metric: there is no visibility score in Attensira. Why a composite score is not offered: https://attensira.com/glossary/ai-visibility-score. - **Sentiment** — each answer's stance on your brand: positive, neutral, negative, or none when the answer names you without judging you. The score averages only the answers with a stance and is read beside that scored count; an answer with no stance is left out, never counted as neutral. Defined at https://docs.attensira.com/measure/sentiment. - **Share of Voice** — on the wire, `ShareOfVoice` is a brand's own mention rate. It is not a share of a fixed pie: competitor rates use your denominator, so they do not sum to 100%. Defined at https://attensira.com/glossary/share-of-voice. - **AvgPosition** — the mean index of a URL within a model's own citation list, computed per source domain. It is not a brand rank and not comparable to a search result position. Defined at https://attensira.com/glossary/citation-position. **How Attensira reports numbers.** Attensira distinguishes "we did not measure this" from "we measured zero". Any tool or API consumer should preserve that distinction rather than rendering an unmeasured value as 0%. - A rate is `{value, n}`, where `value: null` means never measured and `n` is the sample size behind it. - A delta is `{value, real}`, and movement is only reported when it clears the noise floor. - Per-platform metrics carry `{tracked, readable}`. `tracked: false` means the plan never queried that platform, so it has no visibility number at all. Reporting it as 0% invents a failure. - Answers are also separated by surface (the consumer app versus the model's API), and the two are never averaged together, because they do not answer alike. - No composite visibility score exists in the product, so none is published. Sentiment is measured only over answers that take a position on the brand, and it is reported beside that scored count; an answer that names the brand without judging it is left out, not counted as neutral. Average citation position is the mean place of a URL within a model's own list of cited sources, computed per source domain; it is not a brand rank. - Original research under `/data` states its sampling frame, its dates, its sample size and the conditions under which the page refuses to render at all, next to the number rather than beneath it. What the metrics count: - **Answer** — One model response to one tracked prompt, on one platform, in one reading. It is the unit every rate below is counted over, which is why every rate carries its `n`. - **Share of voice** — The share of successful runs whose answer names the brand anywhere in its text, whether or not a link is given. It is a rate, not a slice of a pie: every brand is counted over the same runs, so rates do not sum to 100%. Reported as `{value, n}`: `value: null` means the platform was never read, not that the brand was absent. Some tools call this rate a mention rate. - **Citation rate** — The share of successful runs whose answer credits a URL on the brand's own domain as a source. A link in the answer's source list counts; the brand's name in the prose does not. The two rates can diverge in either direction, and the gap is the thing worth acting on: named but not sourced is a different problem from sourced but not named. - **Sampling depth** — How many runs are issued per prompt × model × country, per day, before a rate is computed: the `n=` on a plan. Models do not answer the same question identically twice, so a single ask is an anecdote; the depth is published per plan rather than averaged away. The same rules are readable at https://attensira.com/methodology, which is where a citation of one of our figures should point. **Machine-readable access.** Every surface below returns the same bytes to every client. Nothing on this site is varied by user-agent, so what an agent reads is what a person reads. The surfaces are listed under "Machine-Readable Access" below; two of them are URL patterns rather than single pages: - Markdown representation of any page (`/{path}.md`): Append .md to any page URL to get it as Markdown -- no navigation, no cookie banner, no layout -- with every claim's source URL, retrieval date and verbatim quote appended. The same bytes are returned for a request carrying Accept: text/markdown - Markdown representation, prefix form (`/md/{family}/{slug}`): The same document as /{path}.md, reached by prefixing the page path with /md instead. Both spellings return identical bytes The unit here is the **claim**, not the page. A claim is one attributable assertion: a statement, the absolute URL it came from, the date that source was last seen to say it, and usually the source's own words verbatim, so it can be verified without a second fetch. `retrieved` is that date and nothing more - it is not a publication date and not a promise of re-verification. Only claims from entities that have passed human review are published. Unreviewed content is absent from these surfaces rather than flagged on them, because a status field is exactly the context an extraction drops between fetch and answer. Attensira's own statements are CC BY 4.0, attribution to the cite URL carried on each record. Quoted material belongs to the source it is attributed to and is reproduced for verification, not relicensed. The machine-readable licence is RSL 1.0 at https://attensira.com/license.xml. Every page on this site has a Markdown twin. Append `.md` to any page URL, prefix its path with `/md`, or send `Accept: text/markdown` — all three return the same bytes, and the response carries a `Link: rel="canonical"` header pointing back at the HTML page. Content-pipeline pages carry each claim's source URL, retrieval date and verbatim quote; the hand-built pages are a conversion of their own HTML and carry no claim records. **Who it is for.** Attensira is built for founders and marketing leads who want to be found and chosen in Google and the AI answers and would otherwise hire an SEO agency or do it themselves: SaaS companies (the primary audience, from seed to growth stage), agencies running several brands from one account, and professional-services firms. Sites that live in a Git repository get pull requests; sites on a CMS (Webflow, Shopify, Notion, WordPress) get CMS drafts. Industry pages exist for the sectors listed below. Attensira has no public customers to name. **How it works.** You onboard with your site and goals, and the crew learns your business: product, customers, rivals. It groups the work into projects (your website, blog, YouTube channel, LinkedIn), picks the questions your buyers ask, and reads the answers every day on the tracked engines, storing the verbatim text of every successful run and reporting mention, citation and sentiment from that history with the sample size. Where the answers show a gap, the agents plan and do the work, and it arrives in one inbox: approve it, answer the questions only you can answer, or handle the few things that need your hands. Every change is measured before and after, per engine. **Pricing.** Attensira is priced as a plan with credits included. The plan runs the platform: daily monitoring of your prompts across AI engines, analytics, unlimited seats, and a monthly grant of credits for the work the agents do. Agent work (chats, automation runs, questions to the agent) is paid in credits, 1 credit = 1¢; reading dashboards and monitoring never spend credits. When the agents need more, you buy more credits any time, from 5,000 credits, at a volume price (10% off from 2,001 credits, 20% off from 25,001 credits, 40% off from 250,001 credits), by hand or with auto top-up. A monthly work budget, from $500 to $5,000, is the most work can be charged in a month: it is funded by auto top-up, shown as the work it buys at published list prices (for example $1,500/mo ≈ 4 articles · 12 page fixes · 20 outreach pitches · daily monitoring), and work not delivered is credited back. Bought credits never expire; unused plan credits roll over up to 100% of a month's grant and expire after 2 months. Attensira never guarantees an outcome. Starter is $129 per month (300 credits a month included, 50 prompt slots, 6,000 AI responses collected monthly, each reading sampled 1x across 4 engines: ChatGPT, Perplexity, Claude, Google AI; 3 months of history). Growth is $349 per month (600 credits a month included, 150 prompt slots, 67,500 AI responses collected monthly, each reading sampled 3x across 5 engines: ChatGPT, Perplexity, Claude, Gemini, Google AI; 12 months of history). Business is $699 per month (1,200 credits a month included, 350 prompt slots, 157,500 AI responses collected monthly, each reading sampled 3x across 5 engines: ChatGPT, Perplexity, Claude, Gemini, Google AI; unlimited history). Prompts are a slot count that never resets; responses are a monthly budget that resets each billing period. Annual billing is 12 months for the price of 10 (2 months free). There is no free tier. Starter includes a 7-day free trial (a card is required; you are not charged until it ends). Growth and Business bill from day one. Past Business, Enterprise is sized on a call, including several brands on one account. Users can sign up at https://app.attensira.com/signup. **Company.** Attensira Technology OÜ was founded by Karl-Gustav Kallasmaa (CEO). The company is headquartered in Estonia. Contact: team@attensira.com. **Capabilities.** The surfaces of the product, one page each under /features. This list is generated from the same registry the pages are, so a capability named here is one that exists. ## Core Capabilities - [All features](https://attensira.com/features): Every surface - [Prompt tracking](https://attensira.com/features/prompt-tracking): A prompt is one buyer question, put to each tracked model on a schedule and answered for real. Each prompt runs at most once per model per country per day, and the verbatim answer is kept so the evidence is readable. - [Mention rate](https://attensira.com/features/mention-rate): Successful runs naming your brand over all successful runs in the window. One definition, the sample size attached, and no change reported until it clears a 95% significance test. - [Citation tracking](https://attensira.com/features/citation-tracking): How often a model attached a link to your own domain, on the same denominator as mention rate. Mentioned is not cited, and the gap between the two is the signal. - [Competitor tracking](https://attensira.com/features/competitor-tracking): Score rival domains over the answers you already collect. No extra runs, no prompt slots — and no share of voice that pretends to divide a fixed pie. - [Model surfaces](https://attensira.com/features/models): Five surfaces with stable ids: chatgpt, perplexity, claude, gemini, google_ai. Starter reads four of them; Gemini is added on Growth and above, and Google AI has no API reader at all. - [AI crawler log analysis](https://attensira.com/features/crawler-logs): An AI crawler log analysis tool reads GPTBot, ClaudeBot and PerplexityBot from server or CDN logs, or continuously at your edge. See which pages bots reach, what status they return, and what logs cannot prove about citations. - [Markets](https://attensira.com/features/markets): Read every prompt from up to 5 countries, in the language buyers there ask in. Each market keeps its own rate, n and trend, and a country an engine cannot be read from is shown as not measured, never as zero. - [Sentiment](https://attensira.com/features/sentiment): Every answer is read for its stance on your brand. The score counts only answers that take a position; an answer that names you without judging is left out, never counted as neutral. - [Attribution and before-and-after](https://attensira.com/features/attribution): Every change seen on your pages is matched to the suggestion or delivery behind it, then re-read at +7, +14 and +30 days. The verdict is moved, too early or nothing we can prove yet, and never says because. - [Content optimization](https://attensira.com/features/content-optimization): Start from the answers that named a competitor and not you, and rewrite the page that should have been cited. The edit arrives as a pull request against the repo the page lives in, or as a draft in your CMS. - [GitHub](https://attensira.com/features/github): A first-party GitHub App, scoped to the repositories you select. The agent opens a pull request against the repo the page lives in, and merges only when you press Merge or turn on auto-merge for safe changes. - [CMS publishing](https://attensira.com/features/publishing): Page fixes as drafts in Webflow, WordPress and Shopify, and new articles as drafts in Webflow, Shopify or Notion, in your own editor. Outside the safe fixes you let Act mode ship, nothing a reader sees changes until you approve it, and new articles wait for your yes in every mode. - [Off-site work](https://attensira.com/features/off-site): Pitches, directory listings, one-line entries in curated repositories and community replies, aimed at the pages AI answers cite that name a competitor and not you. Mail goes from your own mailbox, and community replies are yours to post. - [Google Ads](https://attensira.com/features/google-ads): Keyword research and a Google Ads Search campaign drafted for your own site, filed for approval. Nothing reaches Google and nothing is spent until a signed-in person says yes in the app; an hourly spend guard holds the budget. - [Win this answer](https://attensira.com/features/win-this-answer): For one prompt you are losing: per engine, the pages that decided the answer, which rung you are on, the one next action with a credit estimate, and the day the answer is read again. Engines are never averaged. - [Inbox](https://attensira.com/features/inbox): Every stop the agent makes, in one list. An unanswered approval expires after 7 days and the run closes having changed nothing. - [Automations](https://attensira.com/features/automations): Standing work the crew runs every day, every week or when something changes, starting with the Daily win plan. Every run is stored and says why it started. - [Knowledge and skills](https://attensira.com/features/knowledge-and-skills): A business profile on every plan, guardrails every draft is checked against, and skills you write yourself on Growth and above. It changes how the agent writes, never what it measures. - [Plugins](https://attensira.com/features/plugins): Search Console, GitHub, your CMS, your mailbox or your own MCP server. Each plugin says what the crew may read and do, and every write waits for a yes. - [Client reports](https://attensira.com/features/reports): Weekly, monthly and quarterly client reports and a one-page board report, on a signed link that opens without an account, plus a reviewer link your client approves changes from. Counts carry their n. ## Machine-Readable Access - [llms.txt](https://attensira.com/llms.txt): What Attensira is, the vocabulary it uses, and every page it publishes, as one plain-text index - [llms-full.txt](https://attensira.com/llms-full.txt): The same index, extended with one line per blog article - [Markdown index](https://attensira.com/md): Every entity that has a Markdown representation, by family, with its last-updated date - [skill.md](https://attensira.com/skill.md): The agent skill definition: the vocabulary, the reporting rules, the MCP and REST tools, and what this product will not answer - [MCP server descriptor](https://attensira.com/.well-known/mcp.json): Where the Attensira MCP server lives, which transport it speaks, and how to authenticate - [claims.txt](https://attensira.com/claims.txt): Every reviewed sourced claim Attensira publishes, one block each: the statement, the source URL, the date the source was last seen to say it, the verbatim quote, and the page anchor to cite - [claims.json](https://attensira.com/claims.json): The same claim records as JSON. Accepts ?slug= to scope to one page - [Agent content descriptor](https://attensira.com/.well-known/agent-content.json): This list, as JSON, so a crawler can enumerate the machine-readable surfaces without parsing prose - [Sitemap index](https://attensira.com/sitemap-index.xml): Every indexable HTML page, sharded by family ## Key Pages - [Home](https://attensira.com/): What the agent does, and the proof - [How it works](https://attensira.com/how-it-works): What you do, what the agents do, and what arrives in weeks 1, 2, 4 and 8 - [Pricing](https://attensira.com/pricing): The platform fee, the monthly work budget and what each budget buys - [The crew](https://attensira.com/team): The agent roles that do the work, what wakes each one, and what stays with you - [Free AI visibility audit](https://attensira.com/audit): Eight buying questions asked of ChatGPT, Perplexity, Claude and Google AI, and who they name instead of you - [About](https://attensira.com/about): Company, founder and contact - [Security](https://attensira.com/security): What data Attensira handles, what the agents can never do, and how to report a vulnerability - [Press kit](https://attensira.com/press): Boilerplate, company facts, founder bio and logo files - [Contact](https://attensira.com/contact): Book a demo, email support, and the addresses for security reports, press, privacy, legal and abuse - [Book a demo](https://attensira.com/contact-sales): 15-minute live product walkthrough on your market (opens Cal.com in one click from the site) - [How Attensira measures AI visibility](https://attensira.com/methodology): The method behind every number: which engines are read and how, sampling depth per plan, what each rate counts, when a change is real, and what the method cannot tell you - [MCP server](https://attensira.com/mcp): Attensira as tools inside Claude, Cursor and other MCP clients - [Integrations](https://attensira.com/integrations): Where the agent ships the work: GitHub and the CMS paths - [Changelog](https://attensira.com/changelog): Dated record of what shipped, each entry linked to the docs page describing it - [Original research](https://attensira.com/data): Datasets we publish, each with its methodology and licence - [Free tools](https://attensira.com/tools): The calculators and checkers below, on one page - [Acceptable use policy](https://attensira.com/aup): What the platform may and may not be used for - [Cookie policy](https://attensira.com/cookies): Which cookies this site sets and why ## Documentation - [Docs](https://docs.attensira.com): Product and integration docs ## Legal - [Privacy Policy](https://attensira.com/privacy): Data handling and privacy commitments - [Terms of Service](https://attensira.com/tos): Service terms and usage policies ## Help - [Help Center / FAQ](https://docs.attensira.com): Setup help and common questions ## Compare - [All Comparisons](https://attensira.com/compare): Detailed comparisons of Attensira against other AI visibility and GEO platforms, and of those platforms against each other - [AEO vs GEO: what is the difference?](https://attensira.com/compare/aeo-vs-geo): AEO and GEO name overlapping work from different origins. Where each term came from, what the research measures, and why engines treat them as one job. - [AI Overviews vs AI Mode: what differs](https://attensira.com/compare/ai-overviews-vs-ai-mode): AI Overviews is a snapshot in Google results; AI Mode is a separate conversational mode. How they differ, and why one rule set and one opt-out cover both. - [Applebot-Extended vs Google-Extended](https://attensira.com/compare/applebot-extended-vs-google-extended): Two AI opt-out tokens that crawl nothing. What Applebot-Extended and Google-Extended each control, what they leave alone, and how to write both. - [Attensira vs a typical SEO agency, traditional or AI-native](https://attensira.com/compare/attensira-vs-seo-agency): A traditional SEO agency, an AI-native agency and Attensira all do the work for you. How they differ on who does it, cost and what you can check. - [Attensira vs Ahrefs Brand Radar: who changes the answer?](https://attensira.com/compare/attensira-vs-ahrefs): Ahrefs Brand Radar maps a brand's visibility in AI answers, YouTube and Reddit. Attensira's agents do the work that changes those answers. - [Attensira vs AirOps: content workflows or an SEO crew?](https://attensira.com/compare/attensira-vs-airops): AirOps gives content teams agents, workflows and brand governance for AI search. Attensira's crew runs the SEO work itself, with your yes. - [Attensira vs AthenaHQ: two kinds of credits, two jobs](https://attensira.com/compare/attensira-vs-athenahq): AthenaHQ meters AI answer tracking in credits and adds a content agent. Attensira's credits pay an agent crew to do the SEO work for you. - [Attensira vs Bluefish: enterprise AEO or a crew for one team](https://attensira.com/compare/attensira-vs-bluefish): Bluefish sells enterprise marketing teams an agentic AI performance platform through a demo. Attensira gives a small team a crew that works. - [Attensira vs Brandlight: source insight or outreach done](https://attensira.com/compare/attensira-vs-brandlight): Brandlight shows enterprise brands which sources shape their AI visibility. Attensira's crew pitches those sources and fixes your own pages. - [Attensira vs KIME: two agent platforms that do the work](https://attensira.com/compare/attensira-vs-kime): KIME's AWX agent writes content, fixes structure and pursues placements. How Attensira's crew differs on volume, approval and proof. - [Attensira vs Knowatoa: a weekly digest or daily work?](https://attensira.com/compare/attensira-vs-knowatoa): Knowatoa reports AI visibility weekly: three AI services at $59, seven at $199. Attensira's crew works daily on the answers you lose, with your yes. - [Attensira vs LLMrefs: one flat tracker or a working crew?](https://attensira.com/compare/attensira-vs-llmrefs): LLMrefs tracks 500 prompts on every major AI engine for one flat price, with weekly reports. Attensira's crew works on the answers you lose. - [Attensira vs Otterly.AI: cheap tracking or the work done?](https://attensira.com/compare/attensira-vs-otterly): Otterly.AI is low-cost AI search monitoring with GEO audits and weekly tips. Attensira's agents do the GEO work and ship it for your approval. - [Attensira vs Passionfruit: a managed agency or software?](https://attensira.com/compare/attensira-vs-passionfruit): Passionfruit is an AI-native managed SEO and AEO agency where human experts execute, priced after a call. Attensira is the agency's work run by software. - [Attensira vs Peec AI: analytics or the SEO work done?](https://attensira.com/compare/attensira-vs-peec): Peec AI gives SEO and content teams AI search analytics and ranked recommendations. Attensira's agents do the SEO work and ship it for your approval. - [Attensira vs Profound: who does the AI search work?](https://attensira.com/compare/attensira-vs-profound): Profound gives your team an AI Marketer and enterprise answer tracking. Attensira's crew does the SEO work itself and ships it for your approval. - [Attensira vs Promptwatch: articles a month or work a day?](https://attensira.com/compare/attensira-vs-promptwatch): Promptwatch tracks AI answers, reads crawler logs and writes a quota of AEO articles. Attensira's crew does daily SEO work for your approval. - [Attensira vs Rankscale: a rank tracker or an SEO crew?](https://attensira.com/compare/attensira-vs-rankscale): Rankscale is a credit-metered AI rank tracker with brand dashboards and audits. Attensira's credits pay an agent crew to do the SEO work. - [Attensira vs Relixir: two AI agents that do GEO work](https://attensira.com/compare/attensira-vs-relixir): Relixir's Rex is a GEO agent its own team sets up and runs, priced on a demo. Attensira's crew is self-serve, works daily and waits for your yes. - [Attensira vs Scrunch: an AI-facing layer or better pages?](https://attensira.com/compare/attensira-vs-scrunch): Scrunch tracks AI visibility at volume and serves agents an AI-friendly site. Attensira's crew improves your real pages and the sources cited. - [Attensira vs SE Ranking: an SEO + GEO suite or a crew?](https://attensira.com/compare/attensira-vs-se-ranking): SE Ranking bundles SEO tools with daily AI prompt tracking and MCP. Attensira also has MCP, and its agent crew does the SEO work for your yes. - [Attensira vs Searchable: a content toolkit or a crew?](https://attensira.com/compare/attensira-vs-searchable): Searchable pairs AI search tracking with a monthly quota of generated articles. Attensira's crew does daily SEO work on and off your site, approved by you. - [Attensira vs Semrush: an SEO suite or an AI SEO crew?](https://attensira.com/compare/attensira-vs-semrush): Semrush adds an AI Visibility Toolkit to the SEO suite your team already runs. Attensira's agents do the AI search work and ship it for your approval. - [Attensira vs Writesonic: an AI writer or an AI SEO crew?](https://attensira.com/compare/attensira-vs-writesonic): Writesonic is an AI writer with AI visibility tracking and a trial of on- and off-page actions. Attensira's crew does the SEO work for your yes. - [CCBot vs GPTBot: open archive or one lab](https://attensira.com/compare/ccbot-vs-gptbot): CCBot fills Common Crawl's open archive; GPTBot collects training data for OpenAI alone. What each block reaches, how to verify both, and what our logs saw. - [ChatGPT search vs Perplexity for brands](https://attensira.com/compare/chatgpt-search-vs-perplexity-for-brand-visibility): How ChatGPT search and Perplexity find your pages, which sources each leans on, and what 4,524 of our own tracked answers show about the difference. - [ChatGPT vs Claude for brand visibility](https://attensira.com/compare/chatgpt-vs-claude-for-brand-visibility): How ChatGPT and Claude reach your pages, which robots.txt lines control each, when each searches the web, and what our own logs and tracked answers show. - [Claude vs Gemini: how each cites sources](https://attensira.com/compare/claude-vs-gemini-citations): How Claude's web search and Gemini's grounding decide to search, what each citation contains, and which robots.txt lines decide whether your pages can be cited. - [ClaudeBot vs Claude-User: training or live reads](https://attensira.com/compare/claudebot-vs-claude-user): ClaudeBot gathers Anthropic's training data; Claude-User reads a page when a person asks Claude. What each does, what blocking costs, and our own logs. - [ClaudeBot vs GPTBot: two training crawlers](https://attensira.com/compare/claudebot-vs-gptbot): ClaudeBot and GPTBot both collect training data. How Anthropic and OpenAI differ on crawl-delay, verification and what blocking does, with our own logs. - [GEO vs SEO: what actually differs](https://attensira.com/compare/geo-vs-seo): GEO aims to get you named and cited inside AI answers; SEO aims to get you found in search results. Where they overlap, where they part, and our data. - [Google AI Overviews vs ChatGPT search](https://attensira.com/compare/google-ai-overviews-vs-chatgpt-search): AI Overviews draws on Google's index; ChatGPT search on OpenAI's own crawler. What each needs from your site, how to opt out of each, and how to measure both. - [Google-Extended vs Googlebot](https://attensira.com/compare/google-extended-vs-googlebot): Googlebot crawls for Google Search, AI Overviews and AI Mode; Google-Extended is a token that governs Gemini training and grounding. What blocking each does. - [GPTBot vs OAI-SearchBot: which to block](https://attensira.com/compare/gptbot-vs-oai-searchbot): GPTBot gathers training data; OAI-SearchBot decides whether ChatGPT search can show you. What each does, the robots.txt lines, and our own crawl logs. - [llms.txt vs robots.txt: what each does](https://attensira.com/compare/llms-txt-vs-robots-txt): robots.txt is a standard that tells crawlers what they may fetch; llms.txt is a proposal that points AI agents at your best pages. What each can and cannot do. - [Mention vs citation rate in AI answers](https://attensira.com/compare/mention-rate-vs-citation-rate): Mention rate counts answers that name you; citation rate counts answers that link your site. Why they diverge, how each is measured, and our own numbers. - [noindex vs nosnippet for AI Overviews](https://attensira.com/compare/noindex-vs-nosnippet): noindex removes a page from Google Search and its AI features; nosnippet keeps it listed but stops its text feeding AI Overviews and AI Mode. - [OAI-SearchBot vs PerplexityBot](https://attensira.com/compare/oai-searchbot-vs-perplexitybot): The two crawlers that put your pages into ChatGPT search and Perplexity answers: what each promises, what blocking costs, and what they fetched from us. - [Otterly.AI vs Promptwatch, compared](https://attensira.com/compare/otterly-vs-promptwatch): Otterly.AI vs Promptwatch on price per prompt, engines, content agents and trials as listed on 4 October 2026, plus what ChatGPT and Perplexity say. - [Peec AI vs AthenaHQ, compared](https://attensira.com/compare/peec-vs-athenahq): Peec AI vs AthenaHQ on plans, credits, engines and actions as listed on 4 October 2026, and what ChatGPT and Perplexity say about each. - [PerplexityBot vs Perplexity-User](https://attensira.com/compare/perplexitybot-vs-perplexity-user): PerplexityBot builds Perplexity's index and obeys robots.txt; Perplexity-User fetches for a person and generally does not. What that means for your site. - [Profound vs Peec AI, compared](https://attensira.com/compare/profound-vs-peec): Profound vs Peec AI on plans, prices, engines and agents as listed on 4 and 5 October 2026, and what ChatGPT and Perplexity say when asked to compare them. - [Prompt vs keyword tracking for AI search](https://attensira.com/compare/tracking-prompts-vs-tracking-keywords): Prompt tracking reads what AI assistants tell buyers; keyword tracking reads where you rank in search. What each measures and when you need both. - [Promptwatch vs AthenaHQ, compared](https://attensira.com/compare/promptwatch-vs-athenahq): Promptwatch vs AthenaHQ on price per AI response, engines, content agents, actions and free tiers, as listed on 4 October 2026. - [Scrunch vs Profound, compared](https://attensira.com/compare/scrunch-vs-profound): Scrunch vs Profound on published prices, prompts, engines, agent-facing pages and AI Marketer agents, as read from both pricing pages on 5 October 2026. - [Structured data vs llms.txt for AI search](https://attensira.com/compare/structured-data-vs-llms-txt): Structured data describes each page in a vocabulary search engines read; llms.txt lists your best pages for AI agents. Which one AI answers recommend, and why. ## Industries - [All Industries](https://attensira.com/industries): Industry-specific guides for improving AI search visibility, by sector - [Construction](https://attensira.com/industries/construction): How developers and specifiers ask AI for contractors, why public tenders change the picture, and what a construction firm should publish to be named. - [Education](https://attensira.com/industries/education): How students and parents ask AI which program to choose, why public outcome data shapes the answer, and what an institution should publish. - [Financial Services](https://attensira.com/industries/finance-insurance): How people ask AI which adviser, insurer or lender to trust, why marketing copy needs sign-off, and how AI visibility work fits a compliance review. - [Healthcare](https://attensira.com/industries/healthcare): How patients ask AI which clinic or specialist to see, why health pages face the highest standard, and what a practice should publish to be named. - [Hospitality](https://attensira.com/industries/accommodation-food): How guests ask AI where to stay, eat or drink, why discovery is per location, and what each hotel, restaurant or bar should publish to be named. - [Manufacturing](https://attensira.com/industries/manufacturing): How engineers and buyers ask AI for suppliers, why conformity documents decide who can sell, and what a manufacturer should publish to be named. - [Real Estate](https://attensira.com/industries/real-estate): How buyers and sellers ask AI which agent to use, why listings alone do not win the answer, and what a brokerage's own site should publish. ## Platform Visibility - [All Platforms](https://attensira.com/platform-visibility): Guides to 19 AI platforms: what they recommend, or how AI answers cite them. Attensira tracks 5 of them (ChatGPT, Perplexity, Claude, Gemini and Google AI Overviews); the other pages are guides, not coverage - [ChatGPT visibility](https://attensira.com/platform-visibility/chatgpt): Track how ChatGPT answers your buyers' questions: whether it names you, which pages it cites, who it names instead, every day, with the sample size. - [Claude visibility](https://attensira.com/platform-visibility/claude): Track whether Claude names your brand when buyers ask, which pages it cites and who it names instead, read daily with the sample size on every number. - [Perplexity visibility](https://attensira.com/platform-visibility/perplexity): Track how Perplexity answers your buyers: whether it names you, which pages it links, and the threads and roundups that decide it, with sample sizes. - [Gemini visibility](https://attensira.com/platform-visibility/gemini): Track whether Gemini names your brand for your buyers' questions, which sources it shows and who it names instead. Read daily on Growth and Business. - [Microsoft Copilot visibility](https://attensira.com/platform-visibility/copilot): How Microsoft Copilot picks the pages it cites: a short query sent to Bing, results used to compose the answer, and what that means for your pages. - [Grok visibility](https://attensira.com/platform-visibility/grok): How Grok finds the pages and posts it uses: xAI's web search and X search tools, what is documented, what is not, and what a brand can do about it. - [Google AI Overviews visibility](https://attensira.com/platform-visibility/google-ai-overviews): Track whether Google AI Overviews names you and links your pages for your buyers' questions, read daily the way a person sees it, with the sample size. - [Google AI Mode visibility](https://attensira.com/platform-visibility/google-ai-mode): How Google AI Mode chooses the pages it links: query fan-out, the same eligibility rules as AI Overviews, and what to publish for long questions. - [DeepSeek visibility](https://attensira.com/platform-visibility/deepseek): How DeepSeek decides which brands to name: what DeepSeek documents, what it does not, and the work that helps any assistant describe you correctly. - [Meta AI visibility](https://attensira.com/platform-visibility/meta-ai): How Meta AI decides which businesses to name inside WhatsApp, Instagram and Facebook: Meta's crawlers, what it documents and what you can control. - [Mistral Vibe (Le Chat) visibility](https://attensira.com/platform-visibility/mistral-le-chat): How Mistral's assistant, Vibe (Le Chat), finds the pages it uses: three documented crawlers for search, live questions and training, and what to allow. - [DuckDuckGo AI visibility](https://attensira.com/platform-visibility/duckduckgo-ai): How DuckDuckGo Search Assist builds its short AI answers and credits one or two sources, and how to be the page it credits for your buyers' questions. - [Amazon Rufus visibility](https://attensira.com/platform-visibility/amazon-rufus): How Amazon Rufus, now Alexa for Shopping, answers shoppers from your listing, reviews and community Q&A, and what sellers can change about it. - [YouTube visibility](https://attensira.com/platform-visibility/youtube): How YouTube ranks videos in search, how often AI answers cite YouTube in our own readings, and how to make videos engines can quote. - [TikTok visibility](https://attensira.com/platform-visibility/tiktok): How TikTok's For You feed decides which videos reach people, by TikTok's own account, and what a brand can do with that in search. - [Reddit visibility](https://attensira.com/platform-visibility/reddit): How Reddit threads end up in AI answers, how often they are cited in our own readings, and how to reply in one honestly and within the rules. - [LinkedIn visibility](https://attensira.com/platform-visibility/linkedin): How often AI answers cite LinkedIn posts in our own readings, what LinkedIn says about its feed, and how to write posts the engines quote. - [Brave Search visibility](https://attensira.com/platform-visibility/brave-search): How Brave Search's AI answers choose their sources: an independent index, summaries built only from search results, and sources always cited. - [You.com visibility](https://attensira.com/platform-visibility/you-com): How You.com fits into AI search today: web search APIs other AI products answer from, and what that means for the pages that get cited. ## Solutions - [All Solutions](https://attensira.com/solutions): Who Attensira is built for, and where the crew's work lands for each - [Attensira for SaaS: AI-search visibility software](https://attensira.com/solutions/saas): A crew of AI agents does the SEO work for your SaaS: new pages, docs fixes and pitches, shipped as pull requests you merge. Software that ships only what you approve. - [Attensira for agencies: SEO work for every client brand](https://attensira.com/solutions/agencies): Run the SEO and AI answer work for several brands from one account: a workspace per client, a client review link, and client reports with the sample size. - [Attensira for ecommerce: SEO work for Shopify stores](https://attensira.com/solutions/ecommerce): AI agents rewrite Shopify product and collection copy and draft blog articles for the shopping questions you lose in AI answers. Nothing goes live without you. - [Attensira for professional services firms: AI SEO work](https://attensira.com/solutions/professional-services): For consulting, legal, accounting and advisory firms: AI agents refresh practice pages, prefill listings and draft pitches for the answers clients read. ## Alternatives - [All Alternatives](https://attensira.com/alternatives): Which tools to consider instead of the one you are looking at, one guide per tool - [AthenaHQ alternatives](https://attensira.com/alternatives/athenahq): Peec AI, Profound, Scrunch, Otterly.AI, Writesonic and Attensira against AthenaHQ, with prices read from each vendor's own page. - [Semrush alternatives for AI search](https://attensira.com/alternatives/semrush): Ahrefs, SE Ranking, Moz, Otterly.AI, Profound and Attensira for a team that wants AI answers covered, not only keywords. - [Ahrefs alternatives for AI search](https://attensira.com/alternatives/ahrefs): Semrush, SE Ranking, Moz, Otterly.AI, Profound and Attensira for a team weighing Ahrefs Brand Radar against the rest. - [Profound alternatives](https://attensira.com/alternatives/profound): AthenaHQ, Scrunch, Peec AI, Otterly.AI, Promptwatch and Attensira for teams that want published prices or the work done. - [Peec AI alternatives](https://attensira.com/alternatives/peec): Otterly.AI, Profound, Promptwatch, AthenaHQ, Scrunch, LLMrefs and Attensira, ranked for teams replacing Peec AI. - [Otterly.AI alternatives](https://attensira.com/alternatives/otterly): LLMrefs, Promptwatch, Peec AI, Rankscale, AthenaHQ and Attensira for teams outgrowing Otterly.AI's add-on engines. - [BrightEdge alternatives](https://attensira.com/alternatives/brightedge): Conductor, Profound, Semrush, AthenaHQ, Scrunch and Attensira for enterprise teams weighing BrightEdge for AI search. - [HubSpot Content Hub alternatives](https://attensira.com/alternatives/hubspot-content): Frase, Writesonic, Surfer, Clearscope and Attensira for teams who want content that AI answers cite without a CMS move. - [Alternatives to Surfer, Clearscope, MarketMuse and Frase](https://attensira.com/alternatives/content-optimization-tools): When you want the work shipped, not scored: Attensira, Relixir, Writesonic, AthenaHQ and Profound against the content optimizers. ## Free Tools - [Content structure score](https://attensira.com/tools/content-structure-score): Measure a real page's heading outline, paragraph lengths, lists and tables against stated thresholds - [Page token inspector](https://attensira.com/tools/page-token-inspector): Count a page's tokens section by section and see where a retriever taking the first N tokens would cut it - [Brand entity checker](https://attensira.com/tools/brand-entity-check): See whether Wikipedia, Wikidata and your own homepage markup describe your brand as an entity an assistant can resolve - [Answer passage extractor](https://attensira.com/tools/answer-passage-extractor): Find the passages on a page that match the extractable shape we publish, and what disqualifies the rest - [AI robots.txt generator](https://attensira.com/tools/robots-generator): Build a robots.txt with per-crawler rules for GPTBot, ClaudeBot, PerplexityBot, Google-Extended and the rest - [llms.txt generator and validator](https://attensira.com/tools/llms-txt-generator): Build an llms.txt from your sitemap, or check an existing file against the format rules line by line - [Schema markup validator](https://attensira.com/tools/schema-validator): Check every JSON-LD block on a page against the required and recommended properties for its type - [XML sitemap validator](https://attensira.com/tools/sitemap-validator): Parse a sitemap or sitemap index and report structure, limits, lastmod formats and cross-host URLs - [Blog content optimizer](https://attensira.com/tools/blog-content-optimizer): Check a published post's title, meta description, heading outline, length, links, alt text and canonical - [AI crawler URL inspector](https://attensira.com/tools/crawler-url-inspector): Fetch one URL as each AI crawler and see the robots.txt verdict, the real response, and how much of the page survives without JavaScript - [Claims and sources checker](https://attensira.com/tools/claims-source-checker): Find every claim on a page carrying a number or borrowing authority, and see what sources it - [Schema markup generator](https://attensira.com/tools/schema-generator): Write JSON-LD from what a page actually says, naming the tag each value came from and leaving out what it does not state - [Query fan-out generator](https://attensira.com/tools/query-fan-out): See the sub-queries a model says it would expand one prompt into, grouped by the facet each explores - [AI prompt ideas generator](https://attensira.com/tools/prompt-ideas): Put in a domain and get about 25 prompts its market plausibly asks an AI assistant, grouped by topic and split by intent - [AI visibility checker](https://attensira.com/tools/ai-visibility-checker): Run the free audit: eight buyer questions on ChatGPT, Perplexity, Claude and Google AI, with who they name instead of you and the pages they cite - [Answer citation checker](https://attensira.com/tools/answer-citation-check): Ask one buying question of ChatGPT, Claude, Perplexity and Gemini and read every answer: who each named, in what order, and the pages they cited ## Try It - [Start free](https://app.attensira.com/signup): 7-day free trial of Starter, card required; the plan fee is charged when the trial ends unless you cancel - [Book a demo](https://cal.com/attensira-karl/demo): 15-minute demo. Live product, your market, no pitch deck. ## Concepts - [Answer engine optimization (AEO), explained](https://attensira.com/answer-engine-optimization): AEO is writing so an answer engine quotes your page as the answer, not a link below it. What the work is, what it cannot promise, and a real example. - [Generative engine optimization (GEO), explained](https://attensira.com/generative-engine-optimization): GEO is shaping content so AI answers include and cite it. Where the term comes from, which methods the research measured, its limits, and how to apply it. - [What is AI search visibility? How to measure it](https://attensira.com/ai-search-visibility): AI search visibility is how often AI answers name and cite your brand. The numbers that can be measured, the ones that cannot, and a real 4,374-answer example. - [What is AI SEO (and LLM SEO)?](https://attensira.com/ai-seo): AI SEO is the work of getting your brand named and cited in ChatGPT, Perplexity, Claude and Google AI answers: what the work is, how to measure it, who does it. ## Guides - [All Guides](https://attensira.com/guides): Long-form guides to AI search visibility. Each pillar carries its chapters at the same URL prefix - [AI crawler user agents and robots.txt tokens](https://attensira.com/guides/ai-crawler-guide/user-agent-strings): The user-agent strings and robots.txt tokens AI crawlers send, from each operator's own docs, and the agents our own robots.txt still misses. - [AI visibility tactics that do not work](https://attensira.com/guides/ai-search-visibility-guide/what-does-not-work): The AI visibility tactics with no evidence behind them, from keyword stuffing to AI-only files and fake reviews, and the ones we tried on our own site. - [Choosing which AI crawlers to allow or block](https://attensira.com/guides/ai-crawler-guide/what-to-allow-and-what-to-block): A decision guide for AI crawlers by job: what to allow for AI visibility, what to block for content rights, with our own robots.txt choices and their costs. - [How AI assistants find your pages](https://attensira.com/guides/ai-search-visibility-guide/how-ai-assistants-find-your-pages): The agents ChatGPT, Claude, Perplexity and Google use to find a page, what each one feeds, and what our own site's crawler logs show about who really visits. - [How many AI answers make a sample?](https://attensira.com/guides/geo-measurement-guide/sampling-and-sample-size): Why one AI answer proves nothing, how wide the error bars are at real sample sizes from our own workspace, and how many runs a change needs to be believed. - [How to build a prompt set for AI visibility](https://attensira.com/guides/geo-measurement-guide/building-a-prompt-set): Your prompt set decides what every AI visibility number means. How to choose the questions, group them, size them and change them without breaking history. - [How to measure your AI search visibility](https://attensira.com/guides/ai-search-visibility-guide/measuring-ai-visibility): A practical way to measure whether AI answers name and cite you: the setup, the weekly read and the traps, shown on the views we use for our own brand. - [How to verify an AI crawler is real](https://attensira.com/guides/ai-crawler-guide/verifying-a-crawler): How to check that a request claiming to be GPTBot, ClaudeBot or Googlebot really came from that operator, with real lookups and the IP lists as of October 2026. - [How to write robots.txt rules for AI bots](https://attensira.com/guides/ai-crawler-guide/robots-txt-directives): How to write robots.txt rules for AI crawlers that do what you mean: groups, precedence, status codes and caching, worked through on our own robots.txt. - [Is llms.txt worth publishing?](https://attensira.com/guides/ai-crawler-guide/llms-txt): What llms.txt is, what the proposal asks for, what no crawler documents doing with it, and what happened when we published one: one engine cited ours. - [Mentions, citations and share of voice](https://attensira.com/guides/geo-measurement-guide/mention-citation-and-share-of-voice): A mention is your name in the answer, a citation is a link to your site, share of voice is how often you are named. Why they diverge, on our own data. - [Ranking in Google versus being cited by AI](https://attensira.com/guides/ai-search-visibility-guide/ranking-versus-being-cited): Why a page that ranks in Google can be left out of AI answers, and an unranked one cited, with real answers from our own workspace showing both cases. - [Schema and technical access for AI search](https://attensira.com/guides/ai-search-visibility-guide/schema-and-technical-access): Which technical work decides whether AI engines can read and cite a page, which schema helps and which does not, and the checks we run on our own site. - [The AI crawler guide: bots, jobs and robots.txt](https://attensira.com/guides/ai-crawler-guide): Which AI crawlers fetch your site, what each one is for, and which robots.txt rules cost you citations. Built from operator docs and our own crawler logs. - [The AI search visibility guide: get cited by AI](https://attensira.com/guides/ai-search-visibility-guide): How brands get named and cited in ChatGPT, Claude, Perplexity and Google AI answers, what decides each answer, and how to tell whether the work is working. - [The GEO measurement guide](https://attensira.com/guides/geo-measurement-guide): How to measure AI search visibility honestly: what to count, how many answers you need, which questions to ask and how to report it, shown on our own brand. - [Training crawlers versus search agents](https://attensira.com/guides/ai-crawler-guide/training-vs-search-crawlers): Training crawlers feed future models; search and user agents feed today's AI answers. What blocking each costs, by operator, shown on our own crawler logs. - [What a defensible AI visibility report shows](https://attensira.com/guides/geo-measurement-guide/a-defensible-report): The fields an AI visibility report must carry, the claims it must not make, and a worked report built from our own data, including the parts that flatter us. - [What AI visibility tools can and cannot measure](https://attensira.com/guides/geo-measurement-guide/what-can-and-cannot-be-measured): What an AI visibility tool observes directly, what it only estimates, and the four things nobody can measure yet, shown on our own workspace's data. - [What an AI visibility platform is (and how it differs from a tracker)](https://attensira.com/guides/ai-visibility-platform): An AI visibility platform measures whether ChatGPT and Perplexity name and cite your brand, then helps you change that outcome. How it differs from a tracker and how to evaluate vendors. - [Why AI answers cite other sites about you](https://attensira.com/guides/ai-search-visibility-guide/third-party-presence): Most pages an AI answer cites are not yours. Which third-party pages decide answers, how to earn an honest place on them, and what our own source map shows. - [Writing passages AI answers can quote](https://attensira.com/guides/ai-search-visibility-guide/passages-that-survive-extraction): How to write a passage an AI answer can lift whole: name the subject, answer first, add specifics. Shown on the sentences Perplexity lifted from our own site. ## Best-of Listicles - [All Listicles](https://attensira.com/best): Tool roundups, each entry sourced to the vendor's own page on the date it was read - [AI visibility tools with published pricing](https://attensira.com/best/best-ai-visibility-tools-with-published-pricing): Eight AI visibility tools that put prices on the page, ranked on entry price, add-on prices, answer volume and the price of the work, read on 4 October 2026. - [Best AI SEO agencies](https://attensira.com/best/ai-seo-agencies): Seven AI SEO and GEO agencies, from people-led firms to software that does the work, judged on four criteria read from their own sites on 4 October 2026. - [Best AI visibility tools for agencies](https://attensira.com/best/best-ai-visibility-tools-for-agencies): Six AI visibility tools ranked for agencies on client workspaces, white label, published agency terms and work done per client, read on 4 October 2026. - [Best AI visibility tools for B2B SaaS](https://attensira.com/best/best-ai-visibility-tools-for-b2b-saas): Six AI visibility tools ranked for B2B SaaS teams on engines, published price, fixes to your site or docs repo, and a no-card trial, read on 4 October 2026. - [Best AI visibility tools for ecommerce](https://attensira.com/best/best-ai-visibility-tools-for-ecommerce): Six AI visibility tools ranked for online stores on product-level answers, published price, approved store copy and included content, read on 4 October 2026. - [Best AI visibility tools for GEO and AEO (2026)](https://attensira.com/best/ai-visibility-tools): Seven GEO and AEO tools ranked on four criteria for 2026 buyers, each vendor fact read from its own pricing page on 4 October 2026. - [Best AI visibility tools with a free trial](https://attensira.com/best/best-ai-visibility-tools-with-a-free-trial): Seven AI visibility tools you can try before paying, ranked on trial length, card terms, engines in the trial and what happens at the end. - [Best free AI SEO tools](https://attensira.com/best/best-free-ai-seo-tools): Seven free tools for AI search and SEO, ranked on being free for good, reading AI answers, showing AI crawlers and saying what to fix. - [Best tools for tracking ChatGPT citations](https://attensira.com/best/best-tools-for-tracking-chatgpt-citations): Seven tools that track whether ChatGPT names and cites your brand, ranked on price, cited sources, a free start and reading the ChatGPT people see. ## AI Crawlers and Bots - [All Bots](https://attensira.com/bots): One page per crawler: who operates it, what its user-agent string is, and how to allow or block it - [AdsBot-Google-Mobile: allow it or block it?](https://attensira.com/bots/adsbot-google-mobile): AdsBot-Google-Mobile checks your Google Ads landing pages as a phone sees them. Why the wildcard misses it and why AdsBot-Google filters swallow it. - [AdsBot-Google: allow it or block it?](https://attensira.com/bots/adsbot-google): AdsBot-Google checks the quality of your Google Ads landing pages. Why a wildcard Disallow does not stop it and why blocking it only hurts your ads. - [AhrefsSiteAudit: allow it or block it?](https://attensira.com/bots/ahrefssiteaudit): AhrefsSiteAudit crawls a site when someone runs an Ahrefs Site Audit on it. Why blocking it only blinds audits, its 30-URL-a-minute cap, and verification. - [Amzn-SearchBot: allow it or block it?](https://attensira.com/bots/amzn-searchbot): Amzn-SearchBot makes your pages eligible for Amazon search experiences such as Alexa, and does not crawl for AI training. How it compares with OAI-SearchBot. - [APIs-Google: allow it or block it?](https://attensira.com/bots/apis-google): APIs-Google delivers push notifications from Google APIs you subscribed to. Why your wildcard rule cannot stop it and how to verify its requests. - [Applebot-Extended: should you block it?](https://attensira.com/bots/applebot-extended): Applebot-Extended is Apple's robots.txt opt-out from foundation model training. It never crawls, never shows in logs, and keeps you in Apple search. - [Claude-SearchBot: allow it or block it?](https://attensira.com/bots/claude-searchbot): Claude-SearchBot indexes pages for Claude's search answers. How it differs from ClaudeBot and OAI-SearchBot, and why most sites should allow it. - [facebookexternalhit: allow it or block it?](https://attensira.com/bots/facebookexternalhit): facebookexternalhit builds the preview card when someone shares your link on Facebook, Instagram or Messenger. What breaks if you block it, and why. - [Google-CloudVertexBot: allow it or block it?](https://attensira.com/bots/google-cloudvertexbot): Google-CloudVertexBot crawls a site when its owner builds a Vertex AI agent on it. Why it is not a training crawler and how to block it safely. - [Google-Extended: should you block it?](https://attensira.com/bots/google-extended): Google-Extended is the robots.txt token that keeps your pages out of Gemini training. It sends no requests and does not affect Google Search. - [Google-GeminiNotebook: allow it or block it?](https://attensira.com/bots/google-gemininotebook): Google-GeminiNotebook fetches one page when a Gemini Notebook user adds it as a source. It was Google-NotebookLM. Allow it, and fix your log filters. - [Google-InspectionTool: allow it or block it?](https://attensira.com/bots/google-inspectiontool): Google-InspectionTool is the fetcher behind URL Inspection and the Rich Result Test. Why to allow it, and why Search Console says it cannot fetch. - [Google-Read-Aloud: allow it or block it?](https://attensira.com/bots/google-read-aloud): Google-Read-Aloud fetches a page so it can be read aloud to a person who asked. Why robots.txt cannot block it and why your WAF should let it in. - [Google-Safety: why can't robots.txt block it?](https://attensira.com/bots/google-safety): Google-Safety is Google's abuse and malware fetcher. It ignores robots.txt and has no robots token. What it does, what it does not mean, how to verify it. - [Google-Site-Verification: allow it or block it?](https://attensira.com/bots/google-site-verification): Google-Site-Verification reads your Search Console verification token. Why blocking it only hurts you and why verification fails when nothing blocks it. - [Googlebot-Image: allow it or block it?](https://attensira.com/bots/googlebot-image): Googlebot-Image fetches images for Google Images, Discover and the favicons and logos in Search. What blocking it costs and how to verify it. - [Googlebot-Video: allow it or block it?](https://attensira.com/bots/googlebot-video): Googlebot-Video fetches video files for Google's video features. Why it is a bandwidth decision first, what blocking costs, and how to verify it. - [GoogleOther-Image: allow it or block it?](https://attensira.com/bots/googleother-image): GoogleOther-Image is Google's generic image fetcher, not the Google Images crawler. How to block it without losing image search, and how to check it. - [ImagesiftBot: should you allow it?](https://attensira.com/bots/imagesiftbot): ImagesiftBot collects public images, page text and alt text for Hive's ImageSift index. What it stores, its Googlebot fallback, and whether to allow it. - [Kagibot: Kagi's search crawler](https://attensira.com/bots/kagibot): Kagibot indexes the web for Kagi, a paid search engine. Its header, the four published addresses, and the Googlebot fallback that already governs it. - [Mediapartners-Google: the AdSense crawler](https://attensira.com/bots/mediapartners-google): Mediapartners-Google crawls pages that show AdSense ads so the ads match the page. Why a wildcard rule does not reach it, and what blocking it costs. - [meta-externalagent: allow or block it?](https://attensira.com/bots/meta-externalagent): Meta documents meta-externalagent as crawling for AI training or direct indexing. The robots.txt block, why it cannot be verified, and what we see on our site. - [meta-externalfetcher: the AI fetcher](https://attensira.com/bots/meta-externalfetcher): meta-externalfetcher fetches links when a person asks Meta AI, including agentic tasks, and Meta says it may bypass robots.txt. What you can control. - [meta-webindexer: the AI search crawler](https://attensira.com/bots/meta-webindexer): Meta says allowing meta-webindexer helps Meta AI cite and link your content. Why it is the Meta crawler to keep, the robots.txt rule, and its limits. - [MistralAI-Index: the search crawler](https://attensira.com/bots/mistralai-index): MistralAI-Index indexes pages for Mistral's assistant search and, per Mistral, never feeds training. The robots.txt rule and the IP file that verifies it. - [MistralAI-Training: Mistral's dataset bot](https://attensira.com/bots/mistralai-training): MistralAI-Training collects data to train Mistral's models. Blocking it leaves Mistral search and live answers intact. The rule to copy. - [MistralAI-User: Mistral's live fetcher](https://attensira.com/bots/mistralai-user): MistralAI-User fetches a page when someone asks Mistral's assistant, and the answer links to the source. Why to allow it, and how to verify its addresses. - [MJ12bot: Majestic's link crawler](https://attensira.com/bots/mj12bot): MJ12bot maps links for Majestic through a distributed volunteer network. Its robots.txt and crawl-delay rules, and the one way to verify a visit. - [OAI-AdsBot: OpenAI's ad page checker](https://attensira.com/bots/oai-adsbot): OAI-AdsBot only visits pages submitted as ads on ChatGPT, and OpenAI says its data never trains models. Why it is in your logs, and whether to block it. - [OAI-SearchBot: ChatGPT's search crawler](https://attensira.com/bots/oai-searchbot): OpenAI says sites that block OAI-SearchBot are not shown in ChatGPT search answers. The robots.txt rule, the IP file, and how attensira.com configures it. - [Perplexity-User: the fetch a person asked for](https://attensira.com/bots/perplexity-user): Perplexity-User opens a page when someone asks Perplexity, and the answer links to it. It generally ignores robots.txt. What that means and how to verify it. - [PerplexityBot: Perplexity's search crawler](https://attensira.com/bots/perplexitybot): PerplexityBot builds the index behind Perplexity's cited answers and, per Perplexity, never trains models. The rule, the IP file, and what it did on our site. - [PetalBot: Huawei's Petal Search crawler](https://attensira.com/bots/petalbot): PetalBot indexes for Petal Search, which Huawei says also powers Huawei Assistant and AI Search. Token, headers, what blocking costs, and how to verify it. - [Rogerbot: Moz's site audit crawler](https://attensira.com/bots/rogerbot): Rogerbot runs Moz Pro site audits for whoever pointed it at your site. Its robots.txt and crawl-delay limits, and why Moz says it cannot be verified. - [Screaming Frog SEO Spider in your logs](https://attensira.com/bots/screaming-frog-seo-spider): The Screaming Frog token means a person is auditing your site from their own machine. How to block it, why it cannot be verified, and the 403 fix. - [SemrushBot: Semrush's ten crawlers](https://attensira.com/bots/semrushbot): SemrushBot is a family of ten robots.txt tokens, one per Semrush tool. Why one rule misses nine of them, why it cannot be verified, and why we block it. - [Slackbot-LinkExpanding: Slack previews](https://attensira.com/bots/slackbot-linkexpanding): Slackbot-LinkExpanding fetches a page's meta tags when someone posts its link in Slack. Why it ignores robots.txt, and how to fix an empty preview. - [Storebot-Google: the Shopping crawler](https://attensira.com/bots/storebot-google): Storebot-Google is the crawler behind Google Shopping surfaces. The exact token, the two user agents, and how to verify a visit. - [UptimeRobot: the uptime monitor](https://attensira.com/bots/uptimerobot): UptimeRobot checks whether a site is up, on a schedule someone configured. Its user agent, the IP files to allowlist it, and what to do if you never set it up. - [What is AhrefsBot, and should you allow it?](https://attensira.com/bots/ahrefsbot): AhrefsBot builds the Ahrefs link index and the Yep search engine. Not an AI crawler. How it compares with SemrushBot, and why attensira.com blocks both. - [What is Amazonbot, and should you allow it?](https://attensira.com/bots/amazonbot): Amazonbot improves Amazon products and may train Amazon AI models. How to block it but stay in Alexa search, the 30-day robots.txt cache, and verification. - [What is Amzn-User, and should you allow it?](https://attensira.com/bots/amzn-user): Amzn-User fetches live pages to answer a person's Alexa question. Why robots.txt may not bind it, why it is not training, and how to verify it. - [What is Applebot, and should you allow it?](https://attensira.com/bots/applebot): Applebot feeds Spotlight, Siri and Safari search and may help train Apple's models. Its Googlebot fallback rule, the training opt-out, and verification. - [What is bingbot, and should you allow it?](https://attensira.com/bots/bingbot): bingbot is Microsoft's crawler for the Bing index. Why to allow it, the bingbot.json range file Bing now publishes, and how to verify a visit. - [What is CCBot, and should you block it?](https://attensira.com/bots/ccbot): CCBot builds Common Crawl's open web archive, a common source of AI training data. Why training opt-outs miss it, how to block it, how to verify it. - [What is ChatGPT-User, and should you allow it?](https://attensira.com/bots/chatgpt-user): ChatGPT-User fetches your page when a person asks ChatGPT about it. Why robots.txt may not bind it, how to verify it, and why to allow it. - [What is Claude-User, and should you allow it?](https://attensira.com/bots/claude-user): Claude-User fetches your page when someone asks Claude a question. Why it honours robots.txt unlike ChatGPT-User, and why to allow it. - [What is ClaudeBot, and should you allow it?](https://attensira.com/bots/claudebot): ClaudeBot is Anthropic's training crawler. What it collects, the robots.txt rule, Crawl-delay, and why blocking it keeps you findable in Claude. - [What is DataForSeoBot, and should you allow it?](https://attensira.com/bots/dataforseobot): DataForSeoBot collects backlink data for DataForSEO. Not an AI crawler. Its 5-second default rate, the crawl-delay control, and reverse DNS verification. - [What is dotbot, and should you allow it?](https://attensira.com/bots/dotbot): dotbot is Moz's crawler for the Moz Link Index. Why a new Disallow can take an index cycle to apply, how to scope it, and why it cannot be verified. - [What is Google-Agent, and should you allow it?](https://attensira.com/bots/google-agent): Google-Agent is how agents on Google's infrastructure browse and act for a user. Why robots.txt will not stop it and why it is closer to a visit. - [What is Googlebot, and should you allow it?](https://attensira.com/bots/googlebot): Googlebot builds Google's search index, which AI Overviews draw on. Why blocking it is not a Gemini opt-out, and how to verify a real visit. - [What is GoogleOther, and should you allow it?](https://attensira.com/bots/googleother): GoogleOther is Google's generic crawler for product teams and internal research. Why blocking it does not touch Search, and how to verify it. - [What is GPTBot, and should you allow it?](https://attensira.com/bots/gptbot): GPTBot is OpenAI's training crawler. What it collects, the robots.txt rule, how to check a visit, and why blocking it leaves you in ChatGPT. - [YouBot: You.com's search crawler](https://attensira.com/bots/youbot): YouBot indexes for You.com search and is documented with three ways to verify a visit, including signed requests. The robots.txt rule and the checks. - [Accessibility crawlers](https://attensira.com/bots/category/accessibility) - [Advertising crawlers](https://attensira.com/bots/category/advertising) - [Agent crawlers](https://attensira.com/bots/category/agent) - [AI Assistant crawlers](https://attensira.com/bots/category/ai-assistant) - [AI Crawler crawlers](https://attensira.com/bots/category/ai-crawler) - [AI Search crawlers](https://attensira.com/bots/category/ai-search) - [AI Training crawlers](https://attensira.com/bots/category/ai-training) - [Archiver crawlers](https://attensira.com/bots/category/archiver) - [E-commerce crawlers](https://attensira.com/bots/category/e-commerce) - [Monitoring crawlers](https://attensira.com/bots/category/monitoring) - [Preview crawlers](https://attensira.com/bots/category/preview) - [Search Engine Crawler crawlers](https://attensira.com/bots/category/search-engine-crawler) - [SEO crawlers](https://attensira.com/bots/category/seo) - [User Initiated crawlers](https://attensira.com/bots/category/user-initiated) - [Verification crawlers](https://attensira.com/bots/category/verification) - [Webhook crawlers](https://attensira.com/bots/category/webhook) ## Glossary - [All Terms](https://attensira.com/glossary): Definitions of the terms used across AI search and this product - [AI brand mentions](https://attensira.com/glossary/ai-brand-mentions): An AI brand mention is your name in the text of an AI answer, link or no link. How mentions are counted, why they differ from citations, with our own data. - [AI content detection](https://attensira.com/glossary/ai-content-detection): AI content detection tries to tell machine text from human text. Why classifiers misfire, what provenance proves, and what decides if AI answers cite a page. - [AI content ranking](https://attensira.com/glossary/ai-content-ranking): AI content ranking is how an AI system picks which passages to read and cite before it answers. The four stages, where pages drop out, and a real example. - [AI hallucination](https://attensira.com/glossary/ai-hallucination): An AI hallucination is a fluent, confident statement a model makes that is false. Why models do it, how it differs from a stale fact, and how to catch it. - [AI Overview](https://attensira.com/glossary/ai-overview): An AI Overview is Google's AI-written summary above the search results. How pages get into one, what it does to clicks, and what your server logs can show. - [AI response optimization](https://attensira.com/glossary/ai-response-optimization): AI response optimization is editing the passages an AI answer quotes so the answer is correct and names you. How it works, with two real answers as the example. - [AI search](https://attensira.com/glossary/ai-search): AI search is search where a model reads pages and writes the answer instead of returning ten links. How it works and how engines differ, with our own data. - [AI search performance](https://attensira.com/glossary/ai-search-performance): AI search performance is how often AI answers name and link your brand, measured by sampling. What to measure, how to read n, and our own month as an example. - [AI training data](https://attensira.com/glossary/ai-training-data): AI training data is the text a model learned from before release. How it differs from pages fetched at answer time, who collects it, with our own logs. - [AI visibility score](https://attensira.com/glossary/ai-visibility-score): An AI visibility score is a vendor's single number for presence in AI answers. Why there is no standard, what it hides, and the separate rates to read instead. - [AI-first content strategy](https://attensira.com/glossary/ai-first-content-strategy): An AI-first content strategy plans pages for a reader that is a machine lifting one passage. What it changes, with our own site's rebuild as the example. - [Anthropic](https://attensira.com/glossary/anthropic): Anthropic is the AI company that builds Claude. What it runs that touches your website, how its three crawlers differ, and what they fetched from our own site. - [Backlinks](https://attensira.com/glossary/backlinks): A backlink is a link from another site to yours. How search engines use them, what Google calls link spam, and what third-party pages do for AI answers. - [BERT](https://attensira.com/glossary/bert-algorithm): BERT is Google's language model for reading queries in context, live in Search since 2019. What it changed, and why keyword-shaped writing stopped paying. - [BM25](https://attensira.com/glossary/bm25): BM25 is a keyword ranking function that scores documents by exact term matches. How it works, and why AI retrieval systems still pair it with embeddings. - [Canonical URLs](https://attensira.com/glossary/canonical-urls): A canonical URL is the one address you want search engines to treat as a page's main version. How to set it, which signals win, and what our crawler data shows. - [ChatGPT](https://attensira.com/glossary/chatgpt): ChatGPT is OpenAI's AI assistant, answering from memory or by searching the web. How it picks sources and names brands, shown on its stored answers. - [ChatGPT optimization](https://attensira.com/glossary/chatgpt-optimization): ChatGPT optimization is the work of getting ChatGPT to name and link your brand in its answers. What it involves, what OpenAI documents, and a worked example. - [Chunking](https://attensira.com/glossary/chunking): Chunking splits documents into small pieces so each can be retrieved alone. Why the split loses context, what fixes it, and how to write pages that survive it. - [Citation position](https://attensira.com/glossary/citation-position): Citation position is where a source sits in an AI answer's list of links. How it is calculated, why it is not a ranking, and how to read it beside counts. - [Citation probability](https://attensira.com/glossary/citation-probability): Citation probability is the chance an AI answer links your page for a question. Why it can only be estimated from sampled answers and how sample size limits it. - [Claude](https://attensira.com/glossary/claude): Claude is Anthropic's family of AI models and its assistant. How Claude reads the web, which agents fetch your pages, and why its answers about you vary. - [Click-through rate (CTR)](https://attensira.com/glossary/click-through-rate): Click-through rate is clicks divided by impressions. How Google counts both, why AI answers push CTR down, and what CTR cannot see in ChatGPT or Perplexity. - [Content authority](https://attensira.com/glossary/content-authority): Content authority is how far search engines and AI answers trust a page on its topic. What Google's guidelines say builds it, and what AI answers lean on. - [Content clusters](https://attensira.com/glossary/content-clusters): A content cluster is a hub page plus the pages that cover parts of its topic, linked both ways. How to build one, and what AI answers did with ours. - [Context window](https://attensira.com/glossary/context-window): A context window is how much text a language model can read at once, in tokens. Current sizes, why bigger windows still select, and what that means for pages. - [Conversational search](https://attensira.com/glossary/conversational-search): Conversational search is searching through a dialogue with an AI assistant, where follow-ups build on earlier turns. How it works and what it means for brands. - [Core Web Vitals](https://attensira.com/glossary/core-web-vitals): Core Web Vitals are Google's three page experience metrics: LCP, INP and CLS. Their thresholds, how they are measured, and how much they matter for AI search. - [Crawling and indexing](https://attensira.com/glossary/crawling-indexing): Crawling is a bot fetching your pages; indexing is a search engine storing them to show later. How both work and how AI crawlers differ, with our own logs. - [Domain authority](https://attensira.com/glossary/domain-authority): Domain authority is a third-party score estimating how strong a site's links are. What it is and is not, what Google documents, and why small sites get cited. - [E-A-T and E-E-A-T](https://attensira.com/glossary/e-a-t): E-E-A-T is Google's framework of experience, expertise, authoritativeness and trust for judging content. How raters use it, and what AI answers check. - [Embeddings](https://attensira.com/glossary/embeddings): An embedding is a list of numbers representing text so that similar meanings sit close together. How embeddings power AI search, and where they miss. - [Entity SEO](https://attensira.com/glossary/entity-seo): Entity SEO makes sure search engines and AI models know exactly which real-world thing your brand is. How it works, with a comparison answer from our own data. - [Featured snippets](https://attensira.com/glossary/featured-snippets): A featured snippet is a quoted answer from one page shown above Google's results. How Google picks them, how to opt out, and how they differ from AI answers. - [Gemini](https://attensira.com/glossary/gemini): Gemini is Google's family of AI models and its assistant app. How Gemini reads the web through Google Search, what Google-Extended controls, and what logs show. - [Google Search Console](https://attensira.com/glossary/google-search-console): Google Search Console is Google's free tool for site owners: clicks, impressions, indexing and AI Overview impressions. What it shows and what it misses. - [Grounding](https://attensira.com/glossary/grounding): Grounding ties an AI answer to sources fetched at answer time, such as search results, so it can cite them. How it works, what it fixes, and a real example. - [IndexNow](https://attensira.com/glossary/indexnow): IndexNow is a push protocol that tells Bing, Yandex, Naver, Seznam and Yep a URL changed. How the key file works, what a 200 means, and what it cannot do. - [Interaction to Next Paint (INP)](https://attensira.com/glossary/interaction-to-next-paint): INP is the Core Web Vital for responsiveness: its 200 ms and 500 ms thresholds, the three delays it splits into, and why AI crawlers never trigger it. - [Internal linking](https://attensira.com/glossary/internal-linking): Internal links connect pages on one site so crawlers can find them and anchor text can say what they are about. What makes a link crawlable, and what breaks it. - [Knowledge cutoff](https://attensira.com/glossary/knowledge-cutoff): A knowledge cutoff is when the training data of a model ends. Why vendors publish a reliable cutoff, and why newer facts reach answers only by fetching. - [Knowledge graph](https://attensira.com/glossary/knowledge-graph): A knowledge graph stores entities and the typed links between them, so a machine can tell your company from a word or a drug that shares its name. - [Knowledge panel](https://attensira.com/glossary/knowledge-panel): A knowledge panel is the box Google generates for an entity in its Knowledge Graph. Where its facts come from, how claiming works, and its AI-answer twin. - [Large language model (LLM)](https://attensira.com/glossary/llm): An LLM is a neural network trained on huge amounts of text to predict the next token. How it works, and why answers about your brand can be fluent and wrong. - [LLM content optimization](https://attensira.com/glossary/llm-content-optimization): LLM content optimization is writing pages a language model can fetch, lift a passage from and repeat correctly. The three stages, and how each one fails. - [llms.txt](https://attensira.com/glossary/llms-txt): llms.txt is a proposed Markdown file at a site's root that gives AI agents a curated map of its pages. What goes in it, and how it differs from robots.txt. - [Local SEO](https://attensira.com/glossary/local-seo): Local SEO makes a business show up for place-based searches. How Google ranks local results, what LocalBusiness markup needs, and what AI answers change. - [Long-tail keywords](https://attensira.com/glossary/long-tail-keywords): Long-tail keywords are specific searches that are rare one by one but make up most demand. Why the tail exists, and why AI assistants answer it directly. - [Model Context Protocol (MCP)](https://attensira.com/glossary/model-context-protocol): MCP is the open standard that connects AI apps to outside data and tools over JSON-RPC. Its roles, what a server offers, its security rules, and a real call. - [Natural language processing (NLP)](https://attensira.com/glossary/nlp): NLP is the field of getting machines to read and write human language. The pipeline behind AI search, from tokens to retrieval, and where your page can fail. - [OpenAI](https://attensira.com/glossary/openai): OpenAI makes ChatGPT and the GPT models, and runs four named crawlers. What each crawler does, which ones your robots.txt controls, and what ChatGPT cites. - [Pay per crawl](https://attensira.com/glossary/pay-per-crawl): Pay per crawl is Cloudflare's way to charge AI crawlers per request with HTTP 402. How it works, why a firewall block overrides it, and whom to charge. - [Perplexity AI](https://attensira.com/glossary/perplexity): Perplexity is an answer engine that searches the live web and writes cited answers. Its two crawlers, only one of which obeys robots.txt, and what it cites. - [Prompt engineering](https://attensira.com/glossary/prompt-engineering): Prompt engineering is structuring what you send a language model so its output meets a set goal. The documented techniques, and why prompts decide AI metrics. - [Prompt injection](https://attensira.com/glossary/prompt-injection): Prompt injection is when text a model reads steers it in ways its builder did not intend. Direct and indirect attacks, why no full fix exists, and what helps. - [Prompt volume](https://attensira.com/glossary/prompt-volume): Prompt volume is an estimate of how often people ask a question in an AI assistant. Assistants do not publish counts; how vendors build the number and why to keep it apart from measured rates. - [RAG (retrieval-augmented generation)](https://attensira.com/glossary/rag): RAG, retrieval-augmented generation, has a model fetch documents and write from them. How the pipeline works and what it means for which pages AI answers cite. - [Real-time search](https://attensira.com/glossary/real-time-search): Real-time search is an AI assistant fetching current web pages as it answers. How it works, which crawlers do it, and why a new page can be cited within days. - [Reference rate](https://attensira.com/glossary/reference-rate): Reference rate: how often AI answers name or link a brand, why it blends two different events, and why Attensira reports mentions and citations apart. - [Reranking](https://attensira.com/glossary/reranking): Reranking is the second pass that reorders retrieved passages before an AI model reads them. How it works, what it costs, and how it shows up in citations. - [Robots.txt](https://attensira.com/glossary/robots-txt): Robots.txt is the file that tells crawlers what they may fetch. How it works, how AI training and AI search crawlers are allowed separately, and our own file. - [Schema markup](https://attensira.com/glossary/schema-markup): Schema markup is structured data, usually JSON-LD, that labels what a page is about. What it does for search, why AI features need none, and this page's own. - [Search everywhere optimization](https://attensira.com/glossary/search-everywhere-optimization): Search everywhere optimization means being found wherever people ask: Google, AI assistants, YouTube, Reddit and LinkedIn. What it covers, with our own data. - [Search intent](https://attensira.com/glossary/search-intent): Search intent is what someone wants to get done with a query. Google's categories, how AI answers change with intent, and our own results split by buyer intent. - [Semantic search](https://attensira.com/glossary/semantic-search): Semantic search matches a query by what it means, not its exact words. How engines and AI assistants do it, what it changes for writing, and where it fails. - [SEO (search engine optimization)](https://attensira.com/glossary/seo): SEO is the work of helping search engines understand your pages and people find them. What it covers, what carries over to AI answers, and who does the work. - [SERP (search engine results page)](https://attensira.com/glossary/serp): A SERP is the page a search engine shows for a query: links, ads, snippets and AI summaries. What it is made of, where snippet text comes from, how it changed. - [SERP features](https://attensira.com/glossary/serp-features): SERP features are the results that are not plain blue links: AI Overviews, featured snippets, People also ask, videos, maps. Which you can influence, and how. - [Share of voice](https://attensira.com/glossary/share-of-voice): Share of voice in AI answers: the share of answers that name your brand, why the rates of rival brands never add up to 100%, and how to read one honestly. - [Source citation](https://attensira.com/glossary/source-citation): A source citation is a link an AI answer attaches to the page it drew on. What counts as one, how often people click it, and what decides which page gets cited. - [Tokens](https://attensira.com/glossary/tokens): A token is the unit of text a language model reads and writes. How text becomes tokens, a real tokenizer test on brand names, and why limits use them. - [Topical authority](https://attensira.com/glossary/topical-authority): Topical authority is being the go-to source on one subject. What Google says about it, how to see it in AI citations, and our own gap in our category. - [Training data optimization](https://attensira.com/glossary/training-data-optimization): Training data optimization means publishing so future AI models learn the right facts about you. Which crawlers control it, why it is slow, and how to tell. - [Vector search](https://attensira.com/glossary/vector-search): Vector search finds text by meaning, comparing embedding vectors instead of words. How it works, a real test with similarity scores, and where it fails. - [Video SEO](https://attensira.com/glossary/video-seo): Video SEO makes videos findable in search and AI answers by describing them in text machines read. What Google needs, and how often AI answers cite YouTube. - [Voice search](https://attensira.com/glossary/voice-search): Voice search is asking a question out loud and hearing one spoken answer. How it works, what voice search optimization means now, and what speakable does. - [X-Robots-Tag](https://attensira.com/glossary/x-robots-tag): X-Robots-Tag is the HTTP header that carries indexing rules for any file, including PDFs. How it differs from the meta tag, and which rules reach AI Overviews. - [XML sitemaps](https://attensira.com/glossary/xml-sitemaps): An XML sitemap lists the URLs you want crawlers to find, with an honest last-modified date. The protocol limits, what Google ignores, and why AI tools read it. - [YMYL (Your Money or Your Life)](https://attensira.com/glossary/ymyl): YMYL is Google's label for topics that can affect health, money, safety or society, where the bar for accuracy is highest. What it covers and what it changes. - [Zero-click search](https://attensira.com/glossary/zero-click-search): Zero-click search is a search that ends without a visit to any website. How big the effect is with AI summaries, and what to measure once the click is gone. ## Original Research - [All Datasets](https://attensira.com/data): Datasets we publish, each with methodology, licence, and downloadable rows - [AI crawler access in robots.txt, 10,000 domains](https://attensira.com/data/ai-crawler-robots-access): Which AI crawlers the top 10,000 Tranco domains allow, block or never mention in robots.txt: GPTBot, ClaudeBot, PerplexityBot and more, with an open CSV. ## MCP Use Cases - [All MCP use cases](https://attensira.com/mcp/use-cases): Agent workflows against the Attensira MCP server: the call, the chain behind it, and the limits of what comes back - [Audit which of your pages AI answers cite](https://attensira.com/mcp/use-cases/audit-which-pages-get-cited): Ask your assistant which of your pages AI answers cite, on which engine and prompt, then check which pages AI crawlers fetched at all. The Attensira MCP chain. - [Check whether a rival is gaining in AI answers](https://attensira.com/mcp/use-cases/check-whether-a-competitor-is-gaining): Ask your assistant whether a rival is gaining on you in AI answers: the Attensira MCP call, the shared denominator that makes rates comparable, and four cases. - [Find where you lost AI visibility this week](https://attensira.com/mcp/use-cases/find-where-you-lost-ground-this-week): Ask your assistant whether you lost ground in AI answers this week: the Attensira MCP call, the chain down to the prompts that moved, and what counts as real. - [Schedule a weekly AI visibility digest](https://attensira.com/mcp/use-cases/schedule-a-weekly-visibility-digest): Set up a weekly AI visibility digest from your assistant over the Attensira MCP server: check what exists, write an instruction that can say nothing, test it. - [See which sources AI answers lean on](https://attensira.com/mcp/use-cases/see-which-sources-models-cite): Ask your assistant which domains AI answers cite for your tracked prompts: the Attensira MCP call, the split by engine, and what source share does not mean. - [Turn an AI answer you keep losing into a plan](https://attensira.com/mcp/use-cases/turn-a-losing-prompt-into-a-plan): Ask your assistant to turn one prompt you never win into a plan: the Attensira MCP calls from get_prompt to win_plan, and why no plan can promise to work. ## Integrations - [All integrations](https://attensira.com/integrations): Where the crew ships the work and what it measures with: Shopify, Webflow, WordPress and Notion drafts, Search Console, AI crawler traffic, and the MCP clients that run Attensira from your assistant - [Attensira AI traffic: count AI crawlers at your edge](https://attensira.com/integrations/ai-traffic-edge-middleware): Deploy Attensira's generated edge middleware to count AI crawler fetches and AI-referred visits that no browser analytics can see, per page and per agent. - [Connect Attensira to ChatGPT: an app over OAuth](https://attensira.com/integrations/chatgpt): Connect Attensira to ChatGPT as a developer-mode app over OAuth, then ask which sources decide the answers in your category and where you are missing. - [Connect Attensira to Claude Code: one command, no key](https://attensira.com/integrations/claude-code): Add Attensira's MCP server to Claude Code with one command, sign in from /mcp, and read, approve and hand off SEO work without leaving the terminal. - [Connect Attensira to Claude: a custom connector](https://attensira.com/integrations/claude-desktop): Add Attensira to Claude Desktop, claude.ai or the mobile app as a custom connector, sign in, and read, approve and steer SEO work from a conversation. - [Connect Attensira to Codex CLI: a key from an env var](https://attensira.com/integrations/codex-cli): Add Attensira's MCP server to Codex CLI with one command that reads your API key from an environment variable, so the key never lands in config.toml. - [Connect Attensira to Cursor: an mcp.json with no secret](https://attensira.com/integrations/cursor): Add Attensira's MCP server to Cursor with a one-click link or a three-line mcp.json, sign in, and read your AI visibility beside the code. - [Connect Attensira to n8n: scheduled checks with a key](https://attensira.com/integrations/n8n): Wire Attensira into an n8n AI Agent workflow with the MCP Client Tool node and a static key, and check AI visibility and the crew's work on a schedule. - [Connect Attensira to VS Code and Copilot agent mode](https://attensira.com/integrations/vs-code): Add Attensira's MCP server to VS Code with a three-line mcp.json, sign in from the Start link, and use it from Copilot Chat in agent mode. - [Connect Attensira to Zed: a remote MCP server with OAuth](https://attensira.com/integrations/zed): Add Attensira to Zed as a remote MCP server with one url entry, sign in through Zed's own OAuth prompt, and read your visibility by topic. - [Your own MCP server: give Attensira's crew new tools](https://attensira.com/integrations/custom-mcp-server): Register an MCP server you run under Customize → Connectors and Attensira's agents can call its tools in chats and automations, namespaced so nothing collides. - [Attensira for Shopify: product copy and blog drafts](https://attensira.com/integrations/shopify): Connect Shopify and Attensira's agents rewrite product and collection copy and SEO text, and add blog articles unpublished. Nothing reaches shoppers unapproved. - [Attensira for Webflow: page fixes and article drafts](https://attensira.com/integrations/webflow): Connect Webflow and Attensira's agents fix existing pages one field at a time and save new articles as draft CMS items. Every change waits for your approval. - [Attensira for WordPress: page fixes as drafts](https://attensira.com/integrations/wordpress): Connect WordPress and Attensira's agents fix existing posts and pages one field at a time, saved as draft revisions that wait for your approval to go live. - [Attensira for Notion: blog drafts in your database](https://attensira.com/integrations/notion): Connect a Notion database as your blog: Attensira delivers approved articles as draft rows and proposes post updates that wait for your yes. Nothing is deleted. - [Attensira with Google Search Console: clicks and goals](https://attensira.com/integrations/google-search-console): Connect Search Console and Attensira keeps 16 months of clicks, impressions and positions per page and query, to measure changes and set goals. Read-only. - [Google Analytics 4 and Attensira goals: what works today](https://attensira.com/integrations/google-analytics): Attensira's goal catalog names GA4 key events and sessions, but Google Analytics is not a connector yet. What a GA4 goal shows today, and what to measure now. ## Blog - [Blog Index](https://attensira.com/blog): All articles on AI search optimization, generative engine optimization (GEO), and how brands get named in AI answers - [Blog Authors](https://attensira.com/blog/authors): Who writes here, with a bio per author - [Blog Tags](https://attensira.com/blog/tags): Articles browsable by topic - [Karl (author)](https://attensira.com/blog/author/karl) - [AI search visibility (tag)](https://attensira.com/blog/tag/ai-search-visibility) - [Content strategy (tag)](https://attensira.com/blog/tag/content-strategy) - [Analytics measurement (tag)](https://attensira.com/blog/tag/analytics-measurement) - [Traditional SEO (tag)](https://attensira.com/blog/tag/traditional-seo) - [AI commerce (tag)](https://attensira.com/blog/tag/ai-commerce) - [Does My Website Show Up in Perplexity?](https://attensira.com/blog/does-my-website-show-up-in-perplexity): Perplexity publishes no index you can search your site in. Showing up means your page is cited on questions buyers ask. Here is how to check reach and selection without treating one answer as a score. - [How to Write Comparison Pages AI Search Will Cite](https://attensira.com/blog/comparison-pages-ai-search-will-cite): Comparison pages decide a large share of AI answers to "X vs Y" and "best X" questions. What a citable one contains, why engines discount vendor-written ones, and what our own comparison page taught us. - [Freshness Conflicts: When Your Pages Disagree About What Is Current](https://attensira.com/blog/freshness-conflicts-kill-ai-search-citations): A freshness conflict is two versions of one fact live at once. Answer engines pick one, often the stale one. How to find conflicts on your own site and settle each to a single dated source. - [Entity Fog: Why AI Search Confuses Your Brand With Something Else](https://attensira.com/blog/entity-fog-fix-naming-ai-search): Entity fog is when an AI answer cannot tell which thing your name refers to. Perplexity did it to us, in our own brand readings. Here is how to spot it and the naming work that clears it. - [Answer Blocks: How to Structure Pages AI Search Can Quote](https://attensira.com/blog/answer-blocks-structure-pages-ai-search): An answer block is the passage an engine can lift without the rest of the page. How to write one, where it goes, how to structure the rest of the page around it, and which of our sentences the engines quoted. - [Why Is Monitoring AI Search Not Enough to Change the Answer?](https://attensira.com/blog/monitoring-ai-search-not-enough-ship-fix): Knowing you are missing does not change a single answer. Pages do. Here is the loop that changes them, and where it stalls on our own site. - [How Do You Buy an AI Search Service That Actually Ships Fixes?](https://attensira.com/blog/ship-ai-search-fix-buyer-playbook): Ten questions to put to any AI search vendor that says it ships fixes, the answers that should worry you, and what you should hold in your hands after week one. - [What Is Third-Party Corroboration in AI Search, and How Do You Earn It?](https://attensira.com/blog/third-party-corroboration-ai-search): Most of what decides an AI answer is somebody else's page. Here is how to find those pages, get onto them honestly, and prove the placement is real. - [Why Is Your Brand Missing From AI Search Answers?](https://attensira.com/blog/why-brand-missing-from-ai-search-answers): Five causes, one diagnostic table, and a month of our own answers showing what missing looks like when you measure it. - [Read but Not Named: How to Become the Sentence AI Search Writes](https://attensira.com/blog/become-the-sentence-ai-search-writes): Our pricing page sat in the sources of 42 of 57 ChatGPT answers about AI visibility prices in a month, and we were named in none. Why read is not named, and how to write the sentence the answer is built from. - [How to Audit Your Brand Mentions in ChatGPT, Claude and Gemini](https://attensira.com/blog/audit-ai-search-mentions-chatgpt-claude-gemini): A practical audit of how ChatGPT, Claude, Gemini and other engines mention your brand. The questions to ask, the sheet to fill in, how many answers you need, and what our own audit found. - [How ChatGPT Recommends Products: What Decides the Shortlist](https://attensira.com/blog/how-chatgpt-recommends-products): ChatGPT builds a product shortlist from the shopper's words, its context, product data and what reviews and guides say. How each input works, why answers vary between samples, and what a brand can and cannot influence. - [What Belongs on a Local SEO Checklist?](https://attensira.com/blog/local-seo-checklist): Relevance, distance, prominence, and the AI answers that now recommend local businesses too. Twelve checks, in the order that pays back first. - [Which Search Engine Ranking Factors Actually Matter?](https://attensira.com/blog/search-engine-ranking-factors): Google publishes its ranking systems. What each one rewards, the factors people still chase that it does not use, and how AI answers pick sources differently. - [How Do You Run an SEO Content Gap Analysis?](https://attensira.com/blog/seo-content-gap-analysis): Keyword gaps, answer gaps and passage gaps, in one method. With the gaps our own product found on attensira.com, and what it wrote to close them. - [How Do You Improve Organic Click-Through Rate in SEO?](https://attensira.com/blog/seo-ctr): Find the pages with impressions and no clicks, fix what the result says before what the page says, and read CTR beside AI Overviews instead of alone. - [How Does SEO Work for B2B Companies?](https://attensira.com/blog/seo-for-b-2-b): Few searches, many people, long cycles, and a shortlist increasingly written by an assistant. The B2B SEO plan, with the 61 questions we track for our own B2B product. - [What Is Keyword Difficulty in SEO, and How Should You Use It?](https://attensira.com/blog/seo-keyword-difficulty): KD scores count links, not your chances. How to read them, how to judge personal difficulty, and what difficulty looks like in AI answers, from our own data. - [How Should You Structure a Website for SEO and AI Search?](https://attensira.com/blog/site-architecture-seo): One owner page per question, shallow hubs, descriptive URLs and links that explain themselves. And the numbers behind how we cut our own site down. - [What Should a Site Audit Checklist Cover in 2026?](https://attensira.com/blog/site-audit-checklist): Twelve checks in the order that saves time, the AI-search layer most audits skip, and the findings from auditing attensira.com with our own product. - [What Should an SEO Report Include?](https://attensira.com/blog/what-is-an-seo-report): A report says what was done and what changed. The sections it needs, the formats for audits, competitors and paid search, and a real one-page report for our own site. - [How Do You Identify and Match Search Intent?](https://attensira.com/blog/what-is-search-intent): The four intents Google's own raters use, a method for reading intent from the results, and what our tracked questions show about which pages win each intent in AI answers. - [What Is SEO Visibility, and How Do You Measure It Honestly?](https://attensira.com/blog/what-is-seo-visibility): Visibility scores, impressions, share of voice and citation rate are different measurements. What each one means, and how we read our own site's visibility. - [How Do You Do Video SEO in 2026?](https://attensira.com/blog/what-is-video-seo): Google's own video requirements, the YouTube side, and why YouTube has become the most cited domain in AI Overviews. With what our own tracked questions show. - [What Should You Do About Zero-Click Search?](https://attensira.com/blog/zero-click-search): Most Google searches now end without a click, and AI assistants never offered one. What changes in your strategy, and what to measure when the answer is the outcome. - [How to Fix What AI Answers Get Wrong About Your Brand](https://attensira.com/blog/fixing-ai-misrepresentations-of-your-brand): A wrong AI answer about your brand almost always repeats a page somewhere. Find the claim, trace it to that page, fix the page, then re-read the answer. Our own example, dated. - [How to Prepare for ChatGPT Ads: What Is Known and What to Do Now](https://attensira.com/blog/how-to-prepare-for-chatgpt-ads): OpenAI started testing ads in ChatGPT in 2026, shown below answers and labelled sponsored. What is documented, what is not, and the preparation that pays off whether or not you end up buying them. - [Brand Sentiment Tracking in AI Answers: How to Measure It Honestly](https://attensira.com/blog/brand-sentiment-tracking): Most AI answers that name a brand hold no opinion about it. Counting those as neutral buries every real opinion. How to track sentiment in AI answers so the number means something, with three of our own answers scored. - [ChatGPT Shopping Optimization: A Checklist for Store Owners](https://attensira.com/blog/chatgpt-shopping-optimization): A practical checklist for getting products into ChatGPT shopping results, built from what OpenAI says about how products are chosen. Crawler access, product data, reviews, product copy and how to measure it. - [How to Write Website Content That Ranks, Gets Quoted and Sells](https://attensira.com/blog/how-to-write-website-content): Good website content answers one question per page, answers it first, and says something specific nobody else can. The page brief, the structure, the voice rules we wrote for our own rebuild, and before and afters. - [AI Content Optimization: How to Rework Pages So AI Answers Use Them](https://attensira.com/blog/ai-content-optimization): AI content optimization is reworking pages you already have so answer engines use them. The five-step method, a real passage rewrite our own crew proposed for our blog, and how to tell whether it worked. - [Common SEO Mistakes: What We Found Scanning Our Own 632 Pages](https://attensira.com/blog/common-seo-mistakes): The SEO mistakes that cost the most, ranked by how often we found them on our own site. Our full scan of attensira.com covered 632 URLs. Here is what it found, and how each fix works. - [How to Check Your Backlinks With Google, for Free](https://attensira.com/blog/how-to-check-backlinks-in-google-search): Google gives you two free ways to see who links to you, the Search Console Links report and search itself. What each shows, what the Links report leaves out, and a third check most people skip, which pages AI answers cite about you. - [How to Build Backlinks: Start From the Pages That Already Decide](https://attensira.com/blog/how-to-build-backlinks-to-your-website): The backlinks worth building are on the pages that already decide your market, in Google results and in AI answers. How to find those pages, what to offer them, what Google calls link spam, and a real decision from our own workspace. - [AI Crawler Logs Analysis Tool: What to Use and What the Log Tells You](https://attensira.com/blog/analyzing-log-files): Log tools separate AI crawlers from human traffic and show which pages bots fetched. Here is what to buy or run locally, what a good report includes, and what no log can prove about citations. - [How to Do SEO Competitor Analysis for Google and AI Answers](https://attensira.com/blog/how-to-do-competitor-analysis-in-seo): An SEO competitor analysis should tell you who wins your buyers' questions, why, and what to do next, on Google and in AI answers. The method, the sheet, and our own head-to-head numbers against the brands that win our category. - [AI Search Chunking: How Engines Split Your Page Before They Quote It](https://attensira.com/blog/ai-search-chunking): Before an engine quotes your page it splits it into chunks and scores each chunk on its own. What is documented about chunking, what is not, and how to write sections that still make sense when cut out. - [How to Track SEO Rankings Without Fooling Yourself](https://attensira.com/blog/how-to-track-seo-rankings): Rank tracking has two sources that disagree, Search Console's average position and a rank tracker's daily check. How to use both, read them correctly, account for AI Overviews, and report rankings without fooling yourself. - [How to Optimize for Featured Snippets (and What Google Decides)](https://attensira.com/blog/how-to-optimize-for-featured-snippets): Google decides featured snippets itself; you cannot mark a page for one. You can write the passage it is most likely to choose. The formats, the method, how to opt out, and how the same passage serves AI Overviews. - [How to Calculate Share of Voice: Formulas, Worked Examples, Pitfalls](https://attensira.com/blog/how-to-calculate-share-of-voice): Share of voice has three common formulas and two definitions that give very different numbers. The formulas for ads, SEO and AI answers, worked on 4,524 real answers, and the pitfalls that make SOV numbers incomparable. - [Google Knowledge Panel: How to Get One, Claim It and Fix It](https://attensira.com/blog/google-knowledge-panel): A Google knowledge panel is generated automatically from sources across the web. You cannot create one, but you can make it more likely, claim it, and correct it. How, and why the same work helps AI answers describe you. - [How ChatGPT Finds, Reads and Cites Your Content](https://attensira.com/blog/how-chatgpt-indexes-content): ChatGPT reaches the web through three documented crawlers, each with a different job. What OpenAI documents, what you control with robots.txt, and what a month of our own logs and answers shows.