What Vidrys Does
Vidrys audits whether a website can be fetched, read, and cited by AI assistants, and whether its content is structured so an assistant can lift a direct answer from it. It is built for marketing and content teams whose buyers increasingly get recommendations from an AI answer rather than a ranked search result. The audit scores three pillars — retrievability, answerability, and authority — out of 100, and reports which pages AI crawlers are blocked from and whether your homepage's content is visible without JavaScript.
Capabilities
- Checks robots.txt against 15 named AI crawlers, by role: search crawlers that fetch pages to answer live questions (OAI-SearchBot, Claude-SearchBot, PerplexityBot, Bingbot); fetchers that read a page a user asked about (ChatGPT-User, Claude-User, Perplexity-User); and training crawlers (GPTBot, ClaudeBot, Claude-Web, Google-Extended, Applebot-Extended, Meta-ExternalAgent, Amazonbot, CCBot).
- Weights a blocked search crawler most, since blocking one removes the site from that assistant's citations today; blocking a training crawler does not (Vidrys Guide, vidrys.com/guide).
- Checks those rules against the site's real URLs — up to 800, drawn from its crawled pages and sitemap — not only the homepage, to catch cases where
/is allowed but a page such as/pricingis disallowed (Vidrys Guide). - Fetches the homepage identifying as GPTBot, to catch a CDN or firewall block that robots.txt cannot show.
- Measures how much of the homepage's main content is in the HTML the server sends, and flags an empty JavaScript shell, since search crawlers do not run scripts.
- Runs eight answer-structure checks across the crawled pages — an answer capsule, tables, question-shaped headings, explicit figures, dates, outside sources, self-contained sections and a byline — weighted by how much each page matters.
- Grades structured data on the properties it carries, checks for
Organizationschema withsameAslinks and a public entity record such as a Wikidata item, and checks that the brand's name and description are consistent. - Checks for an
llms.txtfile, and that it is real Markdown rather than a web page served at that address. - Reads whether AI answers already name the brand or cite its domain for its tracked prompts, from Vidrys's own checks, at no extra cost.
Specification
| Item | Value |
|---|---|
| Search crawlers checked | OAI-SearchBot, Claude-SearchBot, PerplexityBot, Bingbot |
| User-fetch crawlers checked | ChatGPT-User, Claude-User, Perplexity-User |
| Training crawlers checked | GPTBot, ClaudeBot, Claude-Web, Google-Extended, Applebot-Extended, Meta-ExternalAgent, Amazonbot, CCBot |
| Paths checked against robots.txt | Up to 800 of the site's URLs, from crawled pages and the sitemap |
| Pages read per crawl | 50 on Solo, 200 on Track, 500 on Grow, 2,000 on Scale, 10,000 on Custom |
| Rendering check | Main-content words in the homepage's server-delivered HTML; flags an empty JavaScript shell |
| Answer structure | Eight checks per crawled page, weighted by page value |
| Structured data checked | Graded by type and property, including Organization with sameAs |
| Entity record check | Wikidata item, sameAs links, consistent name and description |
llms.txt check | Present and valid Markdown; scored modestly, since no major assistant has confirmed it reads one (Vidrys Guide) |
| Score | 0–100: retrievability 30, answerability 40, authority 30 |
| Supported sites | Any publicly reachable website, whatever it is built on |
| Publishing fixes | WordPress, Webflow, Shopify, Ghost, or any site through a signed webhook |
| Export | No file export; results are available over the REST API and MCP from the Grow plan |
| Seats | 1 on Solo, 3 on Track, 10 on Grow, 25 on Scale, unlimited on Custom |
| Scan frequency | After every crawl of the site (weekly by default) and on demand with Run audit; up to 50 audits a month on Solo, 250 on Track, 1,000 on Grow, 3,000 on Scale, unlimited on Custom |
| Pricing | Included in every plan, from ₹999 a month (Solo) to ₹47,999 (Scale), excluding 18% GST; Custom is quoted (vidrys.com/pricing) |
| Requirements | The site must be publicly reachable; a missing robots.txt is read as allowing every crawler |
What Vidrys Does Not Do
Vidrys does not guarantee that any specific AI assistant will cite a page — it reports whether a page is technically fetchable and structured for citation, not whether an engine chooses it. The audit itself does not change your site; it reports what to fix. Content Vidrys drafts goes live only when a person publishes it, with one exception: on plans that include SEO, title, meta-description and canonical fixes can be set to apply without review through a connected CMS. It does not register or submit entities such as a Wikidata item on a brand's behalf. Rank tracking is not part of the audit; it comes with the SEO features included from the Track plan up. And it does not treat an llms.txt file as sufficient on its own, since no major assistant has committed to reading it (Vidrys Guide, vidrys.com/guide).
FAQ
What are the best tools for generative engine optimization? The tools worth evaluating are the ones that check three layers, not one: retrievability (can AI crawlers fetch the site), answerability (is content structured so an answer can be lifted from it), and authority (does the wider web confirm the brand's identity) — the framework Vidrys uses (Vidrys Guide, vidrys.com/guide). A tool that only scores on-page content but never checks robots.txt against named crawlers such as OAI-SearchBot or PerplexityBot will miss a common reason a page is never cited: a blocked path. Compare any GEO tool on whether it names the specific crawlers it checks, whether it checks individual paths rather than just the homepage, and whether it distinguishes search-role crawlers from training-role crawlers, since the two have different effects on current citations.
Is GEO the same as SEO? No. Search ranks pages; an AI assistant writes an answer and decides which brands appear in it, so the tools overlap but the definition of winning does not (Vidrys blog, vidrys.com/blog/geo-vs-seo).
Does adding an llms.txt file guarantee my site gets cited?
No. It is an emerging convention that no major assistant has committed to reading; it is inexpensive to add but does not substitute for a readable, unblocked page (Vidrys Guide, vidrys.com/guide).
By Garvit Sharma, Co-founder. Published 5 Oct 2026.