The Lighthouse for AI searchLitmus - The Lighthouse for AI search visibility | Product Hunt

Is your site
visible to AI?

ChatGPT, Perplexity, and AI Overviews only cite what they can parse. Drop in a URL, get a score across 6 dimensions plus a fix brief your AI or dev can act on. No guesswork. No API calls.

  • Acidic0–49
    AI can't read you
  • Neutral50–79
    Partially visible
  • Base80–100
    Cited-ready

Run a reading

One per line or comma-separated. Adds a content gap + comparison analysis vs these sites. Leave blank to skip.

Six reagentsTechnical accessAnswer-first contentAI comprehensionCitation authorityContent freshnessBlog visibility
See it work

Scan it. Fix it or sell it. Watch the score move.

One loop on stage at a time. Click a step, or let it play through.

Step 1 / 6

Drop a URL

Paste a site, hit Run Audit.

the scan
then, pick a path
the payoff

4a and 4b are the same moment, shown two ways: DIY buyer vs. reseller.

Credit packs, not subscriptions

Every credit works the same, however you buy it. Grab one reading when you need it, or stock up so you're never caught without one. Nothing renews, nothing expires, nothing gets billed unless you click buy.

No pack required

The Drop

One dip, one reading. No pack required.

99/ scan

1 scan credit

1=1

One credit runs one scan. Any number of pages, competitor comparisons included.

2 free

Rescan the same site twice for free. The 3rd rescan (then 6th, 9th…) bills like a new scan.

The rest of the shelf, biggest to smallest
The Vial

First real supply, for solo consultants.

10credits
499
The BeakerMost picked

Steady supply, for small agencies.

50credits
1,999
The FlaskSave ~17%

Stock up, lowest cost per reading.

120credits
4,990
The Drum

Scale, with your name on the report.

150credits
4,999
The TankSave ~17%

Agency-scale supply, lowest cost per reading.

600credits
19,990
The ReservoirSave ~17%

For teams running audits at volume.

1800credits
49,990
One scan, two jobs

What a reading actually gets you

Fix it yourself

Close every gap Litmus flags without hiring anyone to do it for you.

  • 01Scan your own site.
  • 02Export the report as a PDF or Markdown file.
  • 03Hand it to your dev team, or paste it into Claude, ChatGPT, or Cursor and let it work the list.

Turn it into billable work

Scan a prospect's site before they've said yes to anything.

  • 01Scan a lead's site. It's public, no permission needed.
  • 02Export a client-ready proposal with your name and company on it, on every plan, even a single credit.
  • 03Show them exactly what's broken. You already did the diagnosis, now bill for the cure.

Watch a reading develop

A 40-second walkthrough of a real audit, start to score.

How it works

No LLM calls. No black box.

AI can only quote what it can reach, parse, and trust. That's a technical gate, separate from how good your writing is. Litmus checks it the same way every time: rule-based, not another model guessing on your behalf.

01

Drop in a URL

Your homepage, or any page. Add up to 3 competitor URLs for a side-by-side.

02

We crawl and render it

Every page, fetched twice: raw HTML and a full render. JavaScript-heavy sites still get scored fairly.

03

Six reagents, one reading

6 dimensions per page, rolled into one site-wide composite score.

04

Export it

For your AI or dev to work through, or as a branded proposal to send a client. Not just a number to stare at.

Questions

Why not just ask ChatGPT or Claude directly?+

Because AI models aren't consistent. Ask the same one the same question twice and you can get two different answers. That's useless as a scoring engine: you'd never know if a lower grade meant your site changed or the model just felt different that time. Litmus instead replicates the consistent, rule-based part of the pipeline: how AI systems crawl, parse, and select sources. Same input, same score, every time.

What is this scoring model based on?+

Our own testing against real AI crawlers, plus published research on AI citation behavior. The field is young and no two sources agree on weighting, so we kept what held up across both and built one scoring system from it. Above 80, the technical gate is solid: keep maintaining it, but shift effort to content instead. It carries more weight in getting cited than any fix beyond 80.

Does this call ChatGPT, Claude, or any LLM to generate my score?+

No. Every check is rule-based: DOM parsing, robots.txt rules, structured-data validation, readability formulas. Same input, same score, every time. No API costs, no rate limits, no black-box randomness.

How is this different from a normal SEO audit tool?+

Traditional SEO optimizes for search-engine ranking signals. Litmus checks a different gate: can AI-answer engines like ChatGPT, Perplexity, and AI Overviews access, parse, and lift citable material from your pages. Some checks overlap with SEO. Several don't.

What counts as a "page" in my max pages setting?+

Any URL Litmus discovers through your sitemap or internal links, up to the limit you set. 200 pages max per scan.

Do rescans cost a credit?+

First 2 rescans of a site: free. Every 3rd scan of that site (3rd, 6th, 9th...) bills like a fresh scan. Rescan after a fix, or monthly, without burning credits every time.

Do you store a copy of my site?+

No. We store your scan results: scores, flagged issues, extracted metadata, tied to your account so you can track progress. Not a mirror of your site's content.

What does one credit cover?+

1 credit = 1 billed site scan, regardless of pages crawled or competitor URLs added for comparison. Free rescans (above) don't touch your balance.

What do I actually get for a credit?+

Not just a score. Every scan includes a full export: a prioritized fix list you can hand straight to your AI or dev team, plus a branded PDF proposal you can send a client. Both come with every plan, even a single credit.

Can I use this to resell audits to clients?+

Yes. Scan any public site, no permission needed, and export a branded proposal with your name and company on it. Several users run Litmus purely as a lead-gen and pitch tool: scan a prospect's site, show them what's broken, sell the fix.

Six reagents

What actually gets measured

Technical Optimization

Can crawlers even get in? Robots.txt, sitemap, structured data.

Answer First Content

Does it answer fast? Clear headings, FAQ blocks, no fluff first.

AI Comprehension

Can AI tell who you are? Consistent naming, plain writing.

Citation Authority

Can it verify you're real? Evidence links, named testimonials.

Content Freshness

Is it current? Visible publish and update dates.

Blog / Content Visibility

Got a blog? The single strongest driver of AI citations we track.

Who's behind this

About Aivirex

Built by AiViREX Innovations LLP, a software studio in Chennai, India. We also build AI-visible websites, AI agents and automations, and custom software end to end.

Litmus vs a manual GEO audit

A manual audit is one consultant's opinion, once. Litmus runs the same rule-based checks every time, so you can track whether a fix actually moved the score.

Where this fits

Content is the key.
We make sure AI can turn it.

Getting cited by AI comes down to your content. Litmus doesn't touch that part, it checks whether AI can even reach the content at all: the six things measured up the strip beside this. Miss enough of them, and the best writing never gets read.

Litmus keeps that gate open, that's the product, today. It doesn't write or judge a word of your content. If you also want the content and strategy behind the gate built for you, that's the harder, bigger-payoff half of GEO/AEO, and it's what Aivirex builds directly.

Technical accessAnswer-first contentAI comprehensionCitation authorityContent freshnessBlog visibilityContentCitedyoursite.com