Everything that decides whether AI cites you
One GEO score, the reasons behind it and the fixes, for every page of your site.
GEO score and checks
One score from 0 to 100 for every page and site, from more than 50 free checks.
6 featuresOne GEO score
A score from 0 to 100 per page and per site, with an A–F grade for sites, from free, repeatable rule checks.
Access
robots.txt rules for AI search and training crawlers, Content-Signal, noindex, nosnippet and noai, and pages that only render with JavaScript.
Quotability
A direct answer near the top, question headings answered right below, lists, tables, steps and short paragraphs.
Attribution and trust
Figures with sources and current years, quotes with sources, authors linked to profiles, and visible dates.
Structured data and entity
Valid JSON-LD, Article fields, Organization schema with its logo and profile links, and FAQ schema answer engines can quote.
Site-wide findings
AI crawlers turned away by a CDN or firewall, llms.txt problems, the brand's name spelled differently, and one question answered on several pages.
Pages
Track the pages that matter, and see what to do next.
4 featuresA dashboard with next steps
Your main website's GEO score, your pages' average, how often AI cites you and the tokens left, with up to four next steps in order.
Up to 100 tracked pages
Labels, tags, notes and pins; import them from a list or a sitemap, and check them all with live progress.
AI reviews and optimizations
Quotable passages and missing entities, then an answer box, FAQ, validated JSON-LD and rewrites. Missing figures are marked, never invented.
History and trends
Every run kept, score trends, what changed since the previous run, ignored issues, Markdown reports and CSV export.
Websites
Analyze a whole site and see what holds it back.
6 featuresOne screen per website
Overview, Content, AI access, Brand, Fixes, AI visibility, Competitors and Reputation, each a page you can link to.
Site analysis
Up to 10 pages with live progress: the site's GEO score, citability, AI crawler access, llms.txt, structured data and the brand entity.
Raise your GEO score
The fixes worth the most points and the score they'd reach, a fix-first plan, and an optional AI-written 90-day GEO plan.
Ready-to-paste fixes
llms.txt, Organization JSON-LD with your brand's profile links, robots.txt rules and Content-Signal lines.
What changed
Any two analyses compared: the score, findings fixed and new, pages, and robots.txt and llms.txt lines.
robots.txt tester
Test any path against every AI crawler, including a draft you paste.
Fixes in your code
An AI model writes the fixes into your repository as a pull request.
4 featuresGitHub or GitLab
Connect with GitHub or GitLab sign-in and pick the repository. Access is stored encrypted; self-managed GitLab works too.
Pick fixes, or fix them all
An AI model reads only the files the findings need, writes exact edits and opens a pull or merge request. It never merges or pushes to your branch.
A bundle for your developer
No repository? Every fix in one free zip: a checklist in order, the ready-made files, a CSV and a brief an AI coding assistant can follow.
Pull requests you can follow
Open, merged or closed, with a nudge to analyze again once the fix is live.
CMS patches
The fixes written into your CMS, previewed and undoable.
4 featuresWordPress, Shopify, Webflow and Ghost
Organization JSON-LD, robots.txt rules, FAQ structured data and the preview image, written where each platform keeps them.
Previewed, confirmed, undoable
Each patch shows its diff first, is written only when you confirm, and can be undone. Nothing someone edited since is overwritten.
Wix, Squarespace and Framer
Guided steps for builders without an API: each fix, where it goes and the text to copy.
Lovable, Bolt and v0
Sites synced to GitHub get fixes that put head tags where crawlers that don't run JavaScript can see them.
Preview images
The card a citation shows, so your source is never a blank tile.
3 featuresMissing images flagged
Pages without a preview image, and a home-page image that's broken or too small.
A card for every page
A clean 1200×630 card with the page's title, your site's name and your color, or your own image.
Upload your own
PNG, JPEG or WebP up to 5 MB, re-encoded without its metadata.
AI visibility
Ask the engines your customers' questions and see who gets cited.
6 featuresFour answer engines
Claude, ChatGPT, Perplexity and Gemini with web search: citation and mention rates, your share of citations, and the sites cited instead.
Why these pages were cited
Each answer matched to the passage of your page it repeats, and what the drawn-on passages have in common. Free.
Ask from a place
Run a check as if the person were in a given city and country.
Unlinked mentions
How often answers name your brand without linking to you, per engine, with the sentence that names you.
Conversational prompts
The multi-sentence messages your audience types to assistants, each with a persona and an intent, and your share of voice by intent.
Content briefs
For questions you aren't cited for: what the cited sources cover, the facts needed, an FAQ and internal links.
AI reputation
What answers say about your brand, and the sources behind it.
2 featuresHow answers describe you
Every sentence naming your brand in the latest check, rated negative, neutral or positive. Free.
Negative-question probe
Each engine asked the questions someone seeding negative associations would target, with the sources behind negative answers and an alert when a new one appears.
Claim check
See which of your claims answer engines can verify.
2 featuresEngines as fact-checkers
Each engine checks up to 5 of your claims on the web and rules each one supported, contradicted or unverified, with its sources.
Confirmed, or only yours
Per claim: confirmed by independent sources, backed only by your own site, disputed or unverified.
Competitor radar and alerts
Where answers cite competitors instead of you.
3 featuresFollow competitors
Up to 10 per website: the answers citing them, their share, the questions they win and on which engines.
Watched sources
The pages cited instead of yours, with the passage answers drew on, so you know exactly what to respond to.
In-app alerts
When a competitor is newly cited, a watched page changes, or a new source backs negative answers. Free.
AI pipeline diagnostics
Where each engine loses your site: discovery, chunking or generation.
3 featuresLayer by layer, per engine
Can its crawlers fetch the pages, do they split into clean passages, and did it find the site but cite others? Free.
Start with the weakest layer
Each engine points to the layer to fix first.
Built-in assistants
Apple Intelligence and Siri, and Microsoft Copilot, get readiness checks for what can be checked from outside a device.
Tools
Quick checks for a passage, a draft or structured data.
3 featuresPassage rewriter
A free checklist of what makes a passage easy to retrieve and quote, and a rewrite that keeps every fact.
Pre-publish GEO check
Paste HTML before it goes live, compared with the live page when it's tracked.
Structured data tools
A structured data tester, and FAQ JSON-LD built from a page's question headings.
Saved reports
Keep a copy of any report, private to you.
2 featuresSave any report
A page run, a site report or an AI visibility check, kept as Markdown. Free.
Private downloads
Each download goes through a private link that works for 10 minutes; nothing is shared.
Plans, spending and accounts
Pay for the AI tokens you use, and see every cost first.
6 featuresPlans priced by AI tokens
Each plan includes AI tokens a month; past them, tokens cost a set price per million, up to the plan's cap.
An exact estimate first
Every AI run shows its token estimate, and what tokens past your allowance would cost, before it starts.
Your choice of model
Claude, GPT-5 or Gemini for reviews, optimizations, plans, briefs, prompts and repository fixes.
A secure account
Sign up with Google, two-factor sign-in, session controls, and a light or dark theme saved on your account.
For developers
A command-line check for CI, and free scheduled re-checks.
Support inside the app
Report a problem with screenshots and get the answer in the app, with a notification. Bugs get a fix proposed for the team to review.
It shows why, then gets it fixed
Monitoring tells you that AI doesn't cite you. Citeplume shows why, and gets it fixed.
It fixes, not just monitors
Findings come with fixes made for you to review: in your code, your CMS or a bundle for your developer. Nothing changes without you.
It shows why you were cited
Not just a count: the passage of your page each answer drew on, and what the cited passages have in common.
It checks what answers say about you
Claim checks have each engine verify your brand's claims on the web, and the reputation probe finds the sources behind negative answers.
It's open about spending
Analyses and GEO checks are free. Anything that uses an AI model shows its exact token estimate first and runs only when you confirm.
See how AI answer engines see your site
Sign up free with Google and analyze your first website. It takes about a minute.