Both tools are free, both need no signup, and both measure readiness, meaning how easy your site is for AI systems to crawl and use. Neither one measures whether an AI actually recommends you. The bottom line: BlueGrid Media's checker is broader and is the only one of the two that grades signals outside your website. Ours is narrower, fully automated, and shows the exact rule behind every pass or fail. If you have 15 minutes, run both, and fix what ours flags first because those are the fixes you can make by editing your own site.
New to the term? What is answer engine optimization covers the basics.
What BlueGrid Media's checker actually is
It is two tools on one page. The instant scan fetches your live page plus robots.txt, llms.txt, and your XML sitemap through a read only proxy, stores nothing, and checks 43 signals in six categories: AI crawler access, AI manifest files, structured data, citation ready content, technical discoverability, and freshness and EEAT. It checks nine AI crawlers by name. BlueGrid warns that password protected or heavily JavaScript rendered sites may return partial results.
The manual deep audit is a separate 11 signal method where you check each item yourself, about 30 seconds each, and enter what you found. The tool then calculates a weighted score out of 100. BlueGrid says the weights come from public research plus its own queries across ChatGPT, Perplexity, Gemini, Claude, and Google AI Overviews, roughly 500 commercial intent queries tracked from December 2025 through May 2026, using the median where studies disagree. The methodology is versioned, and the version we read was v1.0, last verified May 26, 2026. The tool is built for contractors and local service businesses, and BlueGrid reports that most such sites score between 25 and 55 on a first scan.
BlueGrid's 11 weighted signals
| Signal | Points | Where it lives |
|---|---|---|
| Brand mentions and entity authority | 20 | Off your site |
| Google Business Profile health | 15 | Off your site |
| Reviews: diversity, recency, volume | 12 | Off your site |
| Structured data (schema.org) | 10 | On your site |
| Content depth and GEO formatting | 10 | On your site |
| Knowledge Graph and sameAs | 8 | Mixed |
| Freshness signals | 7 | On your site |
| UGC footprint (Reddit, Quora, forums) | 6 | Off your site |
| EEAT and author or license signals | 5 | On your site |
| Earned media and press | 4 | Off your site |
| Technical site health | 3 | On your site |
Adding it up, brand mentions, Google Business Profile, reviews, forum footprint, and earned media account for 57 of the 100 points. By BlueGrid's own weighting, most of an AI visibility score lives off your website. That is the honest strength of their approach, and it is the reason a site only scanner, ours included, cannot be the whole picture.
What our scan actually checks
We read the page you enter plus its robots.txt and llms.txt, then apply eight checks with fixed rules. These are the exact tests, not a description of them.
| Check | Points | The rule it applies |
|---|---|---|
| Schema | 22 | Passes when the page has JSON-LD with a business entity type plus at least one supporting type. Fails on no JSON-LD, unparseable JSON-LD, or no business entity. |
| Server rendered content | 18 | Counts words present in the raw HTML. Under 30 words fails as JavaScript rendered. Under 450 fails as thin. 450 or more passes. |
| Meta tags | 12 | Looks for five signals: a description of 40 or more characters, a canonical link, an Open Graph title, a Twitter card, and a sitemap declared in robots.txt. Four of five passes. |
| Citation worthy content | 12 | Counts concrete facts per 1,000 words: multi digit numbers, dollar amounts, percentages, named entities, and years. Passes at 40 or more signals and 20 or more per 1,000 words. |
| robots.txt | 10 | Fails if the site blocks all bots or blocks any of seven AI search and user fetch bots: OAI-SearchBot, ChatGPT-User, Claude-SearchBot, Claude-User, Claude-Web, PerplexityBot, Perplexity-User. |
| llms.txt | 10 | Passes if a file exists at /llms.txt. |
| Headings | 8 | Requires exactly one H1 and at least three H2 sections. |
| Answer ready content | 8 | Passes with FAQPage or Question schema, or three or more question style headings. |
Two details are worth knowing. First, the robots.txt check looks only at the search and user fetch bots. The emailed report lists all 16 bots and labels each one as AI Search or AI Training, but only the seven search and user fetch bots decide the pass or fail. Nine training crawlers, including GPTBot, ClaudeBot, CCBot, and Google-Extended, are reported without failing the check, because blocking training is a legitimate business decision, while blocking the bots that fetch pages to answer a live question removes you from answers. Second, the content check is a density heuristic. It counts the ingredients of concrete claims, and it cannot judge whether those claims are any good.
After the scan, our report also recommends schema additions by business type. A plumber, for example, is pointed toward service area, service, and FAQ markup, and a restaurant toward menu and opening hours.
Side by side
| Omni scan | BlueGrid instant scan | BlueGrid manual audit | |
|---|---|---|---|
| What it reads | Page, robots.txt, llms.txt | Page, robots.txt, llms.txt, sitemap | Nothing. You answer questions. |
| Scoring | 8 checks, fixed rules, automated | 43 signals in six categories | 11 weighted signals, self reported |
| Weights published | Yes, in this article | Category breakdown shown in the tool | Yes |
| Off site signals | None | None | Brand mentions, reviews, Business Profile, forums, press |
| Bots checked | 16, split into 7 search and 9 training | 9 | Not applicable |
| Repeatable | Same site, same score | Same site, same score | Depends on who fills it in |
| Built for | Any business with a website | Contractors and local service businesses | Contractors and local service businesses |
| What happens after | Fix list by email, managed plans available | Fix list, audit services available | Priority action list |
What neither scanner can prove
Both scores are readiness scores. They count things that should help, and the evidence that any single item causes a citation is thin. In a controlled Ahrefs test, 1,885 pages that added JSON-LD schema showed no meaningful citation increase against 4,000 control pages, though every page in the test was already cited more than 100 times, so the result does not settle whether schema helps a page get noticed at all. We weight schema highest because it is the item we can verify and fix on the page, not because a study proves it moves citations. BlueGrid describes its own weights as derived from observed correlation, and it states that llms.txt is not a confirmed ranking factor for any single engine.
The only way to know whether a fix changed what customers see is to run real questions and count how often you are named. That measurement is noisy. In a study of 2,961 runs across ChatGPT, Claude, and Google's AI features, SparkToro and Gumshoe found less than a 1 in 100 chance that two responses returned the same list of brands. Judge results by the share of many runs that name you, never by one answer.
Which to run, and in what order
- Run our scan on your homepage. It takes about 10 seconds.
- Fix any failing robots.txt or server rendered content check first. A page an AI crawler cannot fetch or read cannot be selected.
- Run BlueGrid's instant scan on the same page to catch what ours does not test, such as freshness and author signals.
- If you are a contractor or local service business, complete BlueGrid's 11 signal manual audit for the off site items, and treat the brand mention, review, and Business Profile items as the long cycle work.
- Write 10 real customer questions, run each five times on ChatGPT, Gemini, and Perplexity, and record how often your business is named. Repeat after each round of fixes.
Questions people ask
Is BlueGrid Media's AI visibility checker really free?
Yes. BlueGrid states there is no signup, no API key, and no paywall on the score or the fix list, and that the scan runs through a read only proxy without storing the site it fetches.
How many signals does the Omni AI visibility scan check?
Eight weighted checks totaling 100 points: schema 22, server rendered content 18, meta tags 12, citation worthy content 12, robots.txt 10, llms.txt 10, headings 8, and answer ready content 8.
Does a high score guarantee AI citations?
No. Both scanners measure readiness. Ahrefs found no citation increase from adding schema to pages already cited more than 100 times, and BlueGrid describes its weights as observed correlations.
Which one should a local service business use?
Use both. BlueGrid's manual audit is built for contractors and local service businesses and grades reviews, Business Profile health, and forum footprint, which our scan does not. Our scan gives a fast, repeatable read of the page itself.