# Maverank: AI visibility tracking and AI readiness audits for brands and agencies > Maverank tracks how AI answer engines talk about a brand, and tells you what to fix so they talk about it more. It monitors named prompts across four AI platforms, scores visibility and share of voice against competitors, and runs a site audit that returns a prioritized, plain-English fix list rather than only a score. This is the full version of [https://maverank.com/llms.txt](https://maverank.com/llms.txt). It is generated from the same source of truth the product runs on: plans and prices come from the plan catalog that gates the application, the platform list from the engine catalog that queries them, and the readiness model from the check definitions that score a site. Nothing here is hand-copied marketing text, so nothing here can quietly go stale. Last generated at deploy time. Canonical origin: https://maverank.com ## What Maverank is Maverank is an AI visibility and AI readiness platform. Two questions, one product: 1. **Visibility.** When a buyer asks an AI assistant for a recommendation in your category, does it name your brand? In what position, with what sentiment, against which competitors, citing which sources? 2. **Readiness.** Can AI systems reach, parse, trust and quote your site at all? Readiness measures the inputs that cause visibility, so it moves before the outcome does. A brand can have low visibility and high readiness (fixes have landed, results are coming) or high visibility and low readiness (currently cited, but fragile). Reporting both is the point. The output of a readiness run is a prioritized fix list written in plain English, ordered by expected impact, ready to hand to a content team or a client. Most tools in this category stop at a score. ## Who it is for - **Agencies** managing a roster of client brands, who need per-client dashboards, white-label reports and something billable to sell: the fixes. - **In-house marketing teams** tracking their own brand and its competitors. - **Consultants and founders** tracking one brand. Read-only client viewers are free and unlimited on every plan, so handing a client their own live dashboard never costs a seat. ## The AI platforms tracked Maverank queries 4 platforms: - ChatGPT - Claude - Gemini - Google AI Overviews Every plan includes all of them, trial included. Platforms are not sold as add-ons and there is no per-platform surcharge. Grok, Perplexity and DeepSeek are not currently tracked; whether Perplexity's *crawler* can reach a site is a separate matter and is measured by the readiness audit. ## How tracking works A **project** is one brand. Adding a project scans the site and runs category searches and AI queries to detect competitors and the prompts buyers actually ask; you review and edit both before tracking starts. A **tracked prompt** is a buyer question run against all platforms on your refresh schedule, for example "best CRM for real estate teams". Your allowance is workspace-wide, not per project, so you decide how to spread it across the roster. Allowances are counted in prompts, never in credits. Per project and per platform, Maverank reports visibility score, share of voice, mention rate, average position and sentiment, which competitors were named instead, and which sources the engines cited. ## The AI Readiness model Every check belongs to one of 5 pillars. Pillars are weighted because they contribute unequally to AI citation behavior: - **Access**, 20% of the score, 7 checks. - **Rendering**, 20% of the score, 1 checks. - **Discoverability**, 20% of the score, 9 checks. - **Entity & structure**, 20% of the score, 6 checks. - **Citation readiness**, 20% of the score, 13 checks. The overall score is 0 to 100, banded as Strong (80+), Moderate (60 to 79), Needs work (40 to 59) and At risk (below 40). 3 of the 36 checks are **gating**: failing one caps the overall score no matter how well the rest scored. A site whose primary copy needs JavaScript to appear, or that blocks the crawlers behind the answer engines, has no readiness worth reporting. A check that could not be measured is reported as **unmeasured** rather than as a failure, and the run's confidence is lowered accordingly. Guessing would be worse than admitting the gap. ### Every check and what it takes to pass - `a1_ai_crawler_access` (Access, gating): robots.txt allows the search/retrieval crawlers OAI-SearchBot, ClaudeBot and PerplexityBot. Training-only or model-control agents are reported separately as informational policy choices. Whether retrieval crawlers can actually fetch the page is a13_crawler_fetchability. - `a14_search_crawler_access` (Access): robots.txt allows Googlebot and Bingbot to fetch the homepage. Google AI Overviews and AI Mode depend on Google Search indexing; Microsoft Copilot discovery depends on Bing's index. - `a13_crawler_fetchability` (Access): Requested with an AI crawler's User-Agent, the homepage returns 2xx and the real HTML: no 403, no bot-management challenge page. robots.txt permission does not establish this on its own, because the WAF or CDN in front of the site never reads robots.txt. - `a2_homepage_reachability` (Access, gating): The homepage responds without errors and time-to-first-byte stays under ~1.6 seconds. - `a3_renders_without_js` (Rendering, gating): At least 80% of the homepage's primary text is readable in the server-rendered HTML, without executing JavaScript. - `a17_key_page_access` (Access): At least 80% of the bounded important-page sample is reachable, has useful server-rendered text and remains eligible for search snippets. - `a4_sitemap` (Discoverability): An XML sitemap exists, is referenced from robots.txt, and carries real, recent dates. - `a5_crawl_traps` (Discoverability): At least 80% of crawled URLs are clean or point at the clean version via rel=canonical. - `a18_hreflang_consistency` (Discoverability): When international variants exist, hreflang codes are valid, every set self-references, has one x-default, uses HTTPS and returns links between sampled equivalents. - `a6_llms_txt` (Discoverability): /llms.txt provides a concise machine-readable site guide; an optional /llms-full.txt companion can provide expanded content. - `a16_indexnow` (Discoverability): The site exposes a valid same-host IndexNow key file and an explicitly authorized test submission is accepted by the IndexNow endpoint. - `a7_indexable` (Access): The page is indexable: no noindex directive in a robots or googlebot meta tag (or the X-Robots-Tag header), so search-index-fed AI engines like Google AI Overviews, Perplexity and Copilot can surface it. - `a15_snippet_eligibility` (Citation readiness): The homepage has no nosnippet or max-snippet:0 directive in its robots meta tags or X-Robots-Tag headers. Google requires snippet eligibility for supporting links in AI Overviews and AI Mode. - `a9_link_headers` (Discoverability): When the site publishes a machine-readable API, service or schema, the homepage response points to it with an appropriate Link relation. - `a10_api_catalog` (Discoverability): For brands that publish an API: /.well-known/api-catalog returns a valid RFC 9727 linkset document listing them. Brands with no public API have nothing to publish here. - `a11_markdown_negotiation` (Discoverability): The homepage answers a request for Accept: text/markdown with a text/markdown Content-Type rather than HTML. - `a12_mcp_server_card` (Discoverability): For software and API brands running an MCP server: /.well-known/mcp/server-card.json (or a recognized variant path) returns valid JSON carrying a serverInfo block. Brands with no MCP server have nothing to declare. - `a8_content_signals` (Access): robots.txt declares no restrictive content signals (no search=no, ai-input=no or ai-train=no) and carries no per-bot Disallow rules for AI crawlers (GPTBot, ClaudeBot, PerplexityBot, Google-Extended, meta-externalagent, CCBot). Declaring nothing at all also passes: undeclared signals read as unrestricted. - `b1_org_schema` (Entity & structure): The homepage carries an appropriate identity entity such as Organization, LocalBusiness or Person and fills at least 4 of 5 portable identity signals. - `b2_key_page_schema` (Entity & structure): At least 80% of key pages carry JSON-LD appropriate to what the page represents, such as Article, Service, Product or Event. - `b3_entity_definition` (Entity & structure): Asked cold, AI describes the brand's category and offering correctly, which follows from a homepage one-liner that states the category in plain words. - `b5_heading_hierarchy` (Entity & structure): Each key page has one clear H1 and descriptive subheadings where the page is long enough to need sections. - `b6_meta_quality` (Entity & structure): Key pages carry a of 10-70 characters and a meta description of 50-160 characters. - `c1_cited_source_presence` (Citation readiness): You are present on the domains that account for at least 80% of citations among the top 15 sources AI cites in your category (weighted by citation count; the big sources matter most). - `c2_review_presence` (Citation readiness): A discoverable profile exists on at least two independent review surfaces that serve this kind of business. - `c3_backlink_authority` (Citation readiness): Backlink authority reaches the top tier for your category. Many distinct referring domains, not raw link volume, is what moves it. - `c4_knowledge_graph` (Entity & structure): When the organization is independently notable, an existing Wikidata item claims its official website. Wikipedia is never required. - `c5_freshness` (Citation readiness): For sites that publish articles, documentation or release information, the newest indexable content is at most ~40 days old. - `c6_eeat_signals` (Citation readiness): The site explains who is responsible for it and provides an appropriate way to make contact. Editorial authorship is assessed separately. - `c7_editorial_attribution` (Citation readiness): When editorial pages exist, each sampled article exposes a named author and a machine-readable publication or modification date. - `d1_answer_shaped` (Citation readiness): Key pages open with a one-sentence definition and include FAQ blocks or comparison tables: formats engines can quote verbatim. - `d2_stats_density` (Citation readiness): Core pages carry concrete, verifiable facts appropriate to the site rather than relying only on adjectives. - `d3_topic_coverage` (Citation readiness): At least 80% of your tracked topics have a dedicated, in-depth page. - `d4_pricing_info` (Citation readiness): A commercial site provides the decision-stage information buyers need; the appropriate format may be pricing, a menu, service fees or product details. - `d6_comparison_content` (Citation readiness): For competitive software, ecommerce and professional-service categories, honest comparison or alternatives content covers the main choices. - `d5_readability` (Citation readiness): Key pages read scannably: short paragraphs, bulleted lists and descriptive subheadings throughout. ## Standards and signals measured The readiness audit measures the agent-facing standards most tools ignore: - robots.txt search/retrieval access (OAI-SearchBot, ClaudeBot, PerplexityBot) - AI crawler fetchability (whether a WAF or CDN answers OAI-SearchBot's User-Agent with a challenge or a 403 despite what robots.txt allows) - Content-Signal (search, ai-input, ai-train) - llms.txt and llms-full.txt - Server-rendered HTML (whether primary copy survives without JavaScript) - Schema.org structured data (Organization, FAQPage, SoftwareApplication) - api-catalog (RFC 9727, /.well-known/api-catalog, Linkset) - MCP server cards - Markdown content negotiation - HTTP Link headers for agent discovery Adoption of several of these is genuinely early. They are reported as low-weight opportunities rather than as failures, because marking a site red for not implementing a specification almost nobody implements would be misleading. ## Plans and pricing Annual billing is 10 months for 12. Prices are in US dollars, per month. ### Solo: $59/month, $590/year For consultants and founders tracking their own brand. - 50 tracked prompts, workspace-wide - 3 projects - 2 seats, plus unlimited free client viewers - Refresh: weekly, daily - Full AI Readiness fix list - 50 tracked prompts - 3 projects - All 4 AI platforms - Daily refresh - 2 seats ### Studio: $149/month, $1490/year For in-house teams and growing agencies. - 200 tracked prompts, workspace-wide - 10 projects - 5 seats, plus unlimited free client viewers - Refresh: weekly, daily - Full AI Readiness fix list - 200 tracked prompts - 10 projects - All 4 AI platforms - Daily refresh - 5 seats - White-label & scheduled reports - API access & MCP server ### Agency: $399/month, $3990/year For teams and agencies managing many brands at scale. - 600 tracked prompts, workspace-wide - 30 projects - 15 seats, plus unlimited free client viewers - Refresh: weekly, daily, manual - Full AI Readiness fix list - 600 tracked prompts - 30 projects - All 4 AI platforms - On-demand refresh - 15 seats - White-label & scheduled reports - API access & MCP server ### Enterprise Unlimited projects, SSO and custom contracts are a conversation rather than a plan. Contact hello@maverank.com. ### Trial 7 days, no credit card. Covers 1 project and 25 tracked prompts across all four platforms, with a preview of the AI Readiness fix list. ## What Maverank is not - Maverank is not a Google rank tracker. It measures AI answers, not SERP positions. - Maverank does not sell AI platforms as add-ons. All four are included on every plan, trial included. - Maverank does not meter in credits. Allowances are counted in tracked prompts, workspace-wide. - Maverank does not charge for client access. Read-only client viewers are free and unlimited on every plan. ## Integrations - **REST API.** Projects, dashboards, competitive matrices, citation sources, readiness fix lists and on-demand scans. Team API keys, 1000 requests/hour. Available from the Studio plan. Documentation: https://maverank.com/api/docs/ - **Hosted MCP server.** A Model Context Protocol endpoint an AI agent connects to with a team API key, with nothing to install. Documentation: https://maverank.com/api/docs/mcp/ - **White-label reporting.** PDF exports use the agency's name and uploaded logo instead of Maverank branding. Scheduled report delivery and read-only client dashboards are also available from the Studio plan. ## Company Maverank's entity has two layers, and both are true: - **Brand.** Maverank is a TechDocs Studio product. TechDocs Studio (https://techdocs.studio) is the agency and products hub the family is published under: Maverank, Biel.ai and PushFeedback. It is a brand, not a registered company, so it is never named as the operator of the service. - **Company.** Volcanic Labs S.L. is the Spanish company that legally operates Maverank. It is the entity named on the privacy policy and the terms of service, and it is the entity that acquired v1. - Registered address: Plaza de Galicia, Local 7, 38612, Santa Cruz de Tenerife, Spain. - Founder: David Garcia, who founded both. - Contact: hello@maverank.com. - Maverank began as an SEO audit tool ("Maverank v1"), which Volcanic Labs acquired and then rebuilt for AI visibility. v1 is discontinued and its accounts were migrated to the current service. ### Related properties Every domain in the family, with its role: the brand hub, the sibling products published under it, and the founder. - [TechDocs Studio](https://techdocs.studio): The brand and products hub Maverank is published under. - [Biel.ai](https://biel.ai): Sibling product under the TechDocs Studio brand: AI search and chat for documentation sites. - [PushFeedback](https://pushfeedback.com): Sibling product under the TechDocs Studio brand: feedback widget for docs sites. - [David Garcia](https://davidgarcia.dev): Founder's site. - [GitHub](https://github.com/TechDocsStudio): The engineering org the product is built in. ## Pages - [Maverank](https://maverank.com/): What the product does, and who it is for. - [AI visibility for agencies](https://maverank.com/agencies): Running AI visibility as a billable service: white-label reports, per-client dashboards, free client viewers. - [Pricing](https://maverank.com/pricing): Every plan, what each limit means, and what the trial includes. - [About Maverank](https://maverank.com/about): What Maverank is, who it is for, where it came from, and what it is part of. - [Free AI visibility tools](https://maverank.com/tools): Mechanical AI visibility, crawler and llms.txt checks, plus free planning resources. - [Free AI Visibility Readiness Score](https://maverank.com/tools/ai-visibility-checker): A transparent 0–100 website score based only on public mechanical signals, with no answer-engine or LLM calls. - [Blog](https://maverank.com/blog): Deep dives, guides and original measurement on AI visibility, AI readiness and the standards that decide whether an engine can read a site. - [AEO glossary](https://maverank.com/glossary): Definitions for the answer engine optimization vocabulary: AEO, GEO, AI visibility, llms.txt, Content-Signal, api-catalog, MCP server cards, agent readiness. - [Privacy policy](https://maverank.com/privacy): What is collected, why, and GDPR rights. - [Terms of service](https://maverank.com/terms): The terms governing use of the service. ## Machine-readable endpoints - [https://maverank.com/llms.txt](https://maverank.com/llms.txt): the short index. - [https://maverank.com/robots.txt](https://maverank.com/robots.txt): every AI user agent named and explicitly allowed, with one Content-Signal declaration (search=yes, ai-input=yes, ai-train=yes). - [https://maverank.com/sitemap.xml](https://maverank.com/sitemap.xml): every public URL, with lastmod. - [https://maverank.com/api/docs/](https://maverank.com/api/docs/): REST API reference. - [https://maverank.com/api/docs/mcp/](https://maverank.com/api/docs/mcp/): MCP server reference.