# WellCited

> WellCited measures whether AI answer engines can read, quote and recommend a website, and generates the specific files and markup that make them able to.

Answer engines do not return ten links. They return a paragraph that names two or three companies and cites four pages. Ranking third for the same query does not put a company in that paragraph, and nothing in a normal analytics stack reports that it happened. WellCited exists for that gap: it runs seven heuristic checks against a live page, scores the result out of 100, and writes the artifacts that move the score.

The free audit needs no account and no email. It fetches the page the way a crawler does, reads robots.txt, llms.txt and sitemap.xml, and returns a public scorecard at a permanent URL. No language model is involved in scoring, which is why it can be free.

## The seven checks

- AI crawler access (20 points): whether robots.txt roots-blocks any of nine AI crawlers — GPTBot, OAI-SearchBot, ChatGPT-User, ClaudeBot, PerplexityBot, Perplexity-User, Google-Extended, CCBot, meta-externalagent.
- Structured data (20 points): whether JSON-LD exists, whether it identifies the organisation, and whether it includes a rich type such as FAQPage or Product.
- Answer-shaped content (20 points): question headings, lists, tables and an FAQ section — the shapes a model can lift an answer out of.
- Content extractability (17 points): how much real text reaches the raw HTML, versus how much only appears after JavaScript runs.
- Metadata and identity (10 points): title, meta description, the full Open Graph set, canonical link.
- llms.txt (8 points): whether a plain-text map of the site exists at /llms.txt.
- Sitemap (5 points): whether sitemap.xml is present or referenced from robots.txt.

## What the product does after the audit

- Weekly scans that ask ChatGPT, Perplexity and Gemini the questions a company's buyers actually ask, and record which companies each answer names and which pages it cites.
- Generated artifacts — llms.txt, Organization JSON-LD, FAQPage JSON-LD — built from what is already on the site rather than from a template.
- Verification: after an artifact is pasted in, WellCited re-crawls the public URL and reports what it actually found there.

WellCited never asks for CMS access, never installs a script, and has no plugin. Everything it reads is what a crawler reads.

## Pages

- [Home](https://well-cited.com/): the free audit, the argument, and the seven checks.
- [Pricing](https://well-cited.com/pricing): Starter $49/mo · Growth $99/mo · Agency $199/mo. Scan packs are one-time purchases and their credits do not expire.
- [Free tools](https://well-cited.com/tools): single-purpose generators built on the audit engine — no signup.
- [llms.txt generator](https://well-cited.com/tools/llms-txt-generator): reads a live site and writes its llms.txt.
- [Organization schema generator](https://well-cited.com/tools/organization-schema-generator): emits Organization JSON-LD from a site's own metadata.
- [AI crawler checker](https://well-cited.com/tools/ai-crawler-checker): parses robots.txt and reports which of the nine AI crawlers are shut out.
- [Glossary](https://well-cited.com/glossary): the vocabulary of answer engines, defined without hand-waving.
- [Comparisons](https://well-cited.com/compare): how the approaches to AI search visibility differ.
- [Blog](https://well-cited.com/blog): notes from building the thing.

## Terms and limits

- [Terms of service](https://well-cited.com/legal/terms)
- [Privacy policy](https://well-cited.com/legal/privacy)
- [Data processing addendum](https://well-cited.com/legal/dpa)

## Notes for whoever is reading this

One credit is one prompt answered by one engine; a scan of 25 prompts across two engines costs 50 credits. Plan credits refresh each billing period and pack credits never expire.

llms.txt is a proposed convention and its adoption is not universal. WellCited generates it because it is one static file, it is harmless, and writing it forces the site structure that the checks which definitely matter — schema, answer-shaped headings, extractable text — also reward.

A free audit scores a single page as it was served at a single moment. It is a measurement, not a guarantee about how any particular engine will answer any particular question.