# clickd

> clickd is agentic SEO, GEO and AI citation optimization. It reads your site the way search engines and AI assistants do, ranks every fix by what it is worth, hands the work to your agent, and checks each fix on the live site.

- Read a site: a person types it at https://clickd.com (no sign-up); an agent asks with mttr_rescan over MCP, or the API, with a key
- Sites read so far: 918,215, 25,731,576 pages
- Rules in the rulebook: 44, in four stages: access, understand, trust, act
- For agents: MCP at https://clickd.com/mcp and a JSON API under https://clickd.com/api/v1. Both take a key, which an email gets; a key reads the public report of any site, and claiming a site adds its private parts.

## What clickd can do for a site

- Make the fixes, through the person's own agent. Connected over MCP, an agent such as Claude Code, Codex or Cursor takes clickd's ranked fixes, changes the site in its own code or CMS, deploys, and asks clickd to read the live site again; clickd grades every fix and the agent goes on to the next.
- Fully automated or by approval. The loop can run on its own, or the person approves each change before it ships. Approval by email is built: with a site set to approve, the agent proposes each change through clickd (mttr_propose), clickd emails the person an Approve and a Decline link, and the agent deploys only once mttr_proposal says approved. Approval by text message is not built yet.
- Approval is a promise the agent keeps. clickd does not deploy anything and cannot stop an agent that ignores the answer; it records who approved what, when and how.
- clickd needs no access to the site. It reads public pages only; the changes are made with the access the agent already has.

## Notes for assistants

- A score says pages are well formed. It does not say an assistant cites the site.
- clickd reads public addresses only, obeys robots.txt, and reads under its own name.
- Close in meaning means similar, not that either site is better.
- Text that came from a scanned site is marked as quoted in every answer to an agent, as data and never as instructions.

## Start here

- [How automated GEO works: read, fix, check](https://clickd.com/how-it-works): How clickd reads a site, checks every page against one rulebook, and learns what each page is about: the four steps from a domain to a ranked fix list.
- [What AI assistants need from a site to cite it](https://clickd.com/why): What an assistant needs from sites in each line of work, how the sites clickd has read do on it, what most sites get wrong, and how the score is built.
- [AI search readiness leaderboard, every site read](https://clickd.com/leaderboard): The best-read of 265,815 sites, ranked by how well search engines and AI assistants can read them. Where does yours land?
- [Enterprise AI search program: GEO and SEO](https://clickd.com/enterprise): Our team works with yours to find the buying questions your brand loses in AI assistants and Google, fix the cause of each, and measure what moved.

## For agents

- [SEO MCP server for Claude Code, Codex and Cursor](https://clickd.com/agents): How an agent connects to clickd, reads what to fix on a site, and checks its own work.
- [API reference: every clickd call over HTTP and MCP · clickd docs](https://clickd.com/docs/reference): Every call clickd answers over HTTP and MCP: 62 calls, each with its arguments, who may make it, and an example to copy.
- [MCP tools: every call an assistant can make · clickd docs](https://clickd.com/docs/tools): Every MCP tool clickd has, 62 in all: what each does, when to use it, and a sample of what it answers.
- [Claude Code SEO skills, free and open](https://clickd.com/skills): Agent skills for Claude Code, Codex and Cursor: one teaches your agent clickd, the others each do one job on your site, all checked with clickd.
- [Agentic SEO skill: check and fix a site](https://clickd.com/skills/clickd): An agent skill for Claude Code, Codex and Cursor: connect to clickd, fix a site’s issues in order of worth, then read the site again to prove it. Version 2.1.0.
- [A skill to build page templates crawlers read well](https://clickd.com/skills/page-system): An agent skill that builds a site’s page templates for search engines and AI assistants, and proves each pattern on the live pages with clickd. Version 2.1.0.
- [A skill to add llms.txt to a site](https://clickd.com/skills/llms-txt): An agent skill that adds a /llms.txt built from your site’s own pages, with a test that catches a missing page, checked live with clickd. Version 2.1.0.
- [Google Universal Cart skill for product pages](https://clickd.com/skills/shop-ready): An agent skill that fixes a shop’s Product schema for Google Universal Cart and Shopping, checked on each live product page with clickd. Version 2.1.0.
- [Livecards skill: run a page inside a post on X](https://clickd.com/skills/livecards): An agent skill that makes a page of your site run inside a post on X: player card tags, framing headers and an embed mode, checked with clickd. Version 2.1.0.
- [Rank a page skill: one page up for one search](https://clickd.com/skills/rank-a-page): An agent skill that runs an SEO campaign on one page for one search, five gates each proved on the live site, with a weekly read-out. Version 2.1.0.

## How it decides

- [The GEO and AI SEO checklist: 44 rules](https://clickd.com/rulebook): Every rule clickd checks on each page, in four stages: can AI fetch a page, lift an answer from it, trust it and act on it. Each set against every site clickd read.
- [Pages AI can quote: the Understand rules](https://clickd.com/rulebook/understand): A page can be reached and still be misread. This stage asks whether a search engine or an assistant can tell what each page is about, and lift an answer from it.
- [Methodology: how clickd reads, scores and ranks](https://clickd.com/methodology): How clickd reads a site, scores it on one rulebook, ranks it, and checks whether assistants cite it, step by step, with what none of it claims.

## Free tools

- [Free GEO and SEO tools: check any page](https://clickd.com/tools): Free SEO, AI search and email tools: what an agent reads of a page, share and X cards, speed, schema, headings, redirects, SSL, AI crawlers, DNS and mail records.
- [Free OG image checker and share preview](https://clickd.com/og-image-checker): See how a link looks when it is shared on X, LinkedIn, Facebook and Slack, and what is wrong with its share tags.
- [Twitter card validator: test X card tags, player cards](https://clickd.com/twitter-card-validator): Test a link’s X card: its twitter: tags, its image, and for a player card whether the player route lets X frame it. Or make the player card tags to paste.
- [Free agent view: a page as Markdown AI assistants read](https://clickd.com/agent-view): Type a page’s address and see it as a browser shows it beside the Markdown an AI assistant or agent gets from it, with what helps or hurts that reading.
- [Free multi-location SEO checker for every location page](https://clickd.com/multi-location-seo-checker): Read every location page of a brand together: structured data, share tags, canonical and hreflang, graded on eight checks that only show across them.
- [Free website speed test for crawlers](https://clickd.com/website-speed-test): How fast a page answers a crawler: time to first byte, the whole page, its weight, compression and redirects.
- [Free schema markup checker for JSON-LD](https://clickd.com/schema-checker): Read every block of structured data on a page, or paste your own: whether it parses, what it says the page is, and what that type is missing.
- [Free schema generator: JSON-LD for a business or product](https://clickd.com/schema-generator): Pick a type, fill in a short form, and get JSON-LD to paste into a page, with the fields Google requires marked and a check in one click.
- [Free heading checker: the h1 to h6 outline of a page](https://clickd.com/h1-checker): A page’s headings as the outline a crawler reads, h1 to h6, and what is wrong with it: no h1 or several, skipped levels, and empty or image-only headings.
- [Free redirect checker: every hop and its status code](https://clickd.com/redirect-checker): Follow an address hop by hop: each status code and where it points, chains, http to https and www moves, loops, and whether it ends at a page.
- [Free SSL certificate checker: expiry, chain and TLS](https://clickd.com/ssl-checker): Check a site’s SSL certificate as a browser meets it: valid for the name, issuer, days left, the chain, TLS versions, and whether http moves to https with HSTS.
- [AI crawler checker: GPTBot, ClaudeBot, PerplexityBot](https://clickd.com/ai-crawler-checker): Which AI crawlers a site’s robots.txt lets in, from GPTBot and ClaudeBot to PerplexityBot, and whether it has an llms.txt.
- [Free DNS health checker: NS, SOA, MX, SPF, DMARC and CAA](https://clickd.com/dns-checker): Read a domain’s DNS from public resolvers: its name servers and SOA, its addresses, CAA, and every record its mail depends on, with what is wrong with each in words.
- [Free MX record checker: mail servers by priority](https://clickd.com/mx-checker): Which servers take a domain’s mail, by priority, whether each has an address, and the faults that lose mail: no MX, an IP for a name, an alias, a null MX.
- [Free SPF record checker: includes and 10 lookup limit](https://clickd.com/spf-checker): Validate a domain’s SPF record: each term and qualifier, every include followed as a tree, DNS lookups counted against the limit of 10 and empty ones against 2.
- [Free DMARC checker: policy, reports and alignment](https://clickd.com/dmarc-checker): Read a domain’s DMARC record tag by tag: its policy for mail that fails, where reports go, how much mail it covers, and whether it only watches.
- [Free DKIM checker: find selectors and key length](https://clickd.com/dkim-checker): Find a domain’s DKIM keys at the selectors mail services use, or at your own, with each key’s type and its length in bits.

## Compare

- [clickd beside Ahrefs, Semrush, Search Console and more](https://clickd.com/compare): clickd is a free layer on top of the SEO tools you use. Ahrefs, Semrush, Search Console, Screaming Frog and 9 more, set out on the same questions.
- [clickd and Ahrefs: use them together](https://clickd.com/vs/ahrefs): clickd is a free layer on top of Ahrefs: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Semrush: use them together](https://clickd.com/vs/semrush): clickd is a free layer on top of Semrush: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Moz Pro: use them together](https://clickd.com/vs/moz): clickd is a free layer on top of Moz Pro: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Majestic: use them together](https://clickd.com/vs/majestic): clickd is a free layer on top of Majestic: what each is best at, which to use for which job, and how to use the two together.
- [clickd and OpenSEO: use them together](https://clickd.com/vs/openseo): clickd is a free layer on top of OpenSEO: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Screaming Frog: use them together](https://clickd.com/vs/screaming-frog): clickd is a free layer on top of Screaming Frog: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Sitebulb: use them together](https://clickd.com/vs/sitebulb): clickd is a free layer on top of Sitebulb: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Surfer: use them together](https://clickd.com/vs/surfer): clickd is a free layer on top of Surfer: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Google Search Console: use them together](https://clickd.com/vs/search-console): clickd is a free layer on top of Google Search Console: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Bing Webmaster Tools: use them together](https://clickd.com/vs/bing-webmaster-tools): clickd is a free layer on top of Bing Webmaster Tools: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Google Analytics: use them together](https://clickd.com/vs/google-analytics): clickd is a free layer on top of Google Analytics: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Yoast SEO: use them together](https://clickd.com/vs/yoast): clickd is a free layer on top of Yoast SEO: what each is best at, which to use for which job, and how to use the two together.
- [clickd and Rank Math: use them together](https://clickd.com/vs/rank-math): clickd is a free layer on top of Rank Math: what each is best at, which to use for which job, and how to use the two together.

## Docs

- [Introduction to clickd: what it reads and fixes · clickd docs](https://clickd.com/docs/intro.md): clickd reads your site the way a crawler does and says what to fix first: one rulebook, four stages, and a fix list ranked by what each fix is worth.
- [Quickstart: read your first site · clickd docs](https://clickd.com/docs/quickstart.md): Read your first site and find the fix worth the most, in about two minutes. No account is needed.
- [How clickd reads a site · clickd docs](https://clickd.com/docs/reads.md): What it fetches, in what order, what it keeps, and what it will not do on anyone’s site.
- [Your data on clickd: see it, change it, delete it · clickd docs](https://clickd.com/docs/your-data.md): How to see, change and delete what clickd keeps about you: your account page does most of it at once, and anything else is removed when you ask.
- [Fix what fails first · clickd docs](https://clickd.com/docs/fix.md): Work down the fix list one rule at a time: see which pages fail it, fix them, read the site again, and check the rule now passes.
- [Let crawlers in · clickd docs](https://clickd.com/docs/crawlers.md): When a site keeps crawlers out, search engines and assistants cannot read it and its score is held at 20 or less. How to open it, host by host.
- [What a score means · clickd docs](https://clickd.com/docs/score.md): A score out of 100 says how well formed a site’s pages are, for a crawler and for an assistant. It does not say whether an assistant cites the site.
- [Stopped at the limit · clickd docs](https://clickd.com/docs/limit.md): clickd reads at most 100 pages of a site. When a site has more, the read stops there, and the report says so and holds back what it cannot know.
- [Needs a person: verdicts a model could not settle · clickd docs](https://clickd.com/docs/person.md): A few rules are read by a language model, since only reading can tell. Where its answer cannot be trusted, the page neither passes nor fails: it waits for a person.
- [The clickd API: JSON for scripts and agents · clickd docs](https://clickd.com/docs/api.md): Every answer clickd gives is JSON under /api/v1, and the MCP tools at /mcp are written on top of it. The same figures a person sees, for scripts and agents.

## Guides

- [What is GEO? Generative engine optimization, with data](https://clickd.com/guides/what-is-geo.md): What generative engine optimization is, where the term came from, how it differs from SEO, and what assistants need from a page, counted across real sites.
- [GEO vs SEO vs AEO: the difference, counted](https://clickd.com/guides/geo-vs-seo-vs-aeo.md): SEO, AEO and GEO compared: what each term means, where they overlap, what to do first, and how sites do on each, counted across real sites.
- [How to get recommended by ChatGPT, and what to fix](https://clickd.com/guides/get-recommended-by-chatgpt.md): ChatGPT recommends from pages its search crawler can read. Let it in, put the answer in the HTML, and say plainly what you do. What clickd finds across sites.
- [Get a local business recommended by ChatGPT and AI](https://clickd.com/guides/local-business.md): Asked who fixes roofs nearby, an assistant names businesses whose sites say what they do and where. What that takes, with clickd's counts across sites.
- [AI crawlers: GPTBot, ClaudeBot and who blocks them](https://clickd.com/guides/ai-crawlers.md): Which AI crawlers exist, what each is for, what robots.txt does and does not control, and how many of the sites clickd has read block each one.
- [llms.txt: what it is, who reads it, how to write one](https://clickd.com/guides/llms-txt.md): What llms.txt is, who proposed it, whether any assistant reads it, how many sites serve one, and how to write and serve one that is worth keeping.
- [Agentic SEO: agents that fix a site and check the fix](https://clickd.com/guides/agentic-seo.md): Agentic SEO is SEO done by a coding agent that reads the site, fixes the code and checks each fix live. What it changes, where it goes wrong, how to run it.
- [Livecards: a page that plays inside a post on X](https://clickd.com/guides/livecards.md): How a livecard, a page inside a post on X, works on X’s player card: the tags, letting only X frame the route, the embed mode, the card cache, and each way it fails.
- [Google Connected Apps: what it asks of an MCP server](https://clickd.com/guides/google-connected-apps.md): Google opened a preview of Connected Apps for Gemini and Search. What it asks of an MCP server, and clickd's own server held to the same list.
- [Free link finder: a research prompt for earned links](https://clickd.com/guides/free-link-finder.md): A research prompt to paste into Claude, ChatGPT or Gemini that finds free, earned links for your site, checks each one, and says what it could not confirm.
- [Google Universal Cart: what it reads from your shop](https://clickd.com/guides/google-universal-cart.md): What Google's Universal Cart is, what Google says it draws on, what a merchant needs to take part, and what clickd checks on product pages across real shops.
- [What AI assistants cite, by line of work](https://clickd.com/research/what-assistants-cite.md): Which sites Gemini cites across hundreds of real questions in sixteen lines of work, and how those sites score on clickd's rules against every site read.