What Does One Audit Find That Separate Tools Miss?
One audit finds the problems that only show up when two results disagree. A robots.txt checker can pass while a crawl check fails. A crawl check can pass while no AI answer mentions you. Each tool on its own reports a score. Read together, the pair tells you where the problem actually is.
That is the case for running AI visibility checks as one audit instead of one tool at a time. It is not about speed. An earlier version of this post promised a complete audit in under a minute. Radar does not run that way, and we never measured an audit duration, so this rewrite drops the promise and explains what the audit actually does.
What Are Radar's 13 Tools?
Radar's 13 tools split into 6 technical checks and 7 paid checks: Reddit monitoring and six that read AI answers. The free check runs the 6 technical tools on one domain, once per domain, and does not ask AI engines about your brand. A paid audit runs all 13, and the tools that read AI answers sample ChatGPT, Claude, Gemini and Perplexity.
| Category | Tool | What it checks | Plan |
|---|---|---|---|
| Crawlability | Crawl Check | Whether 17 AI crawlers can fetch your pages | Free |
| Crawlability | Robots.txt | What your rules allow, bot by bot | Free |
| Crawlability | llms.txt | Whether the file exists and is well formed | Free |
| Content | AI Readiness | An overall technical readiness score | Free |
| Content | AEO Audit | How answer-ready a page is | Free |
| Content | Schema Audit | Structured data coverage and errors | Free |
| Citations | Citations | Whether answers name and link you | Paid |
| Citations | Citation Test | Whether answers cite a specific URL | Paid |
| Citations | Prompt SOV | Your share of brand mentions in category answers | Paid |
| Community | Discussion of your brand on Reddit | Paid | |
| Community | Source Influence | Which sources answers draw on in your category | Paid |
| Trust | Hallucination Check | Statements about you that the evidence does not support | Paid |
| Trust | Brand Disambiguation | Whether answers confuse you with another business | Paid |
One Radar audit, as of October 2026
From the tool list in src/types/radar.ts and the Radar pricing page
tools in a full audit
in the free technical check
AI engines sampled by the paid tools
for one full audit
Brand Disambiguation became the 13th tool on 5 June 2026. When this post was first published in April 2026, Radar ran 12 tools, so the original title was ahead of the product.
How Does Radar Run 13 Tools in One Audit?
Radar runs its tools in staged batches. Lighter tools run three at a time. The four provider-heavy tools, which send many requests to AI providers, each run in a batch of their own. Each batch starts only after the previous one has finished, and results appear in the dashboard as each tool completes.
How a full audit runs
Staged batches, as of October 2026
Light tools
grouped into small batches
Heavy tools
AEO, Citations, Source Influence, Prompt SOV
Results
shown in the dashboard
Insights
read from the combined results
The batches exist because the heavy tools fan one audit out into many provider requests. The AEO audit, for example, can run an extra pass that asks Gemini and Claude five judgments each. Running those tools alone lets one finish before the next starts, so a slow or rate-limited provider affects one tool instead of the whole audit. We made the change on 12 August 2026.
We do not publish a duration for a full audit. It depends on the site, the providers and the batch, and we have not measured it in a way we would stand behind.
Why Do Cross-Tool Insights Matter?
Cross-tool insights matter because the most useful findings are disagreements. A tool that scores 85 and a tool that scores 30 are each easy to read. The fact that both are true at once is what tells you where to look, and only an audit that holds both results can say it.
Radar's insight generator reads the results of one audit and names those combinations. Four real examples, with the conditions Radar checks:
| Insight | When Radar shows it | What it points to |
|---|---|---|
| Robots.txt fine but crawl access poor | Robots.txt scores 70 or more, the crawl check scores under 50 | Something other than the rules blocks bots: a CDN or firewall, meta robots or x-robots-tag headers |
| Technically visible but not cited | Crawl and readiness scores are good, mentions are low | Access is not the problem; content and topical depth are |
| Low citation rate is content-driven, not robots-driven | Retrieval bots allowed, two or more training bots blocked, mention rate under 50% | Training-bot blocks are a fair policy choice; the gap is elsewhere |
| Blocking AI browsing bots while citation rate is low | One or more browsing bots blocked, mention rate under 50% | The bots that fetch pages for live answers cannot reach you |
The first row is the clearest case for one audit. A robots.txt checker reads your rules and reports that they allow the bots. A crawl check sends requests as those bots and reports that they fail. Run on different days in different tools, the two results look like noise. Read together, they say: your rules are fine, so look at the server.
“Rules look reasonable but actual access failing. Check CDN/WAF blocks, meta robots, or x-robots-tag headers.”
Radar insight, Robots.txt fine but crawl access poor
Cross-tool insights come with a paid audit, alongside fix threads, unlimited Strategist messages, PDF export, saved history and competitor comparison.
How Does Radar Decide What to Fix First?
As of October 2026, Radar orders every action by its impact and effort, with ties broken the same way every time, so the same audit always gives the same order. It does not yet rank fixes by traffic, because none of its 13 tools measures clicks.
Radar's code includes a ranking formula that weighs the traffic at stake, the share of it a fix could recover, confidence and effort in hours. It switches on for any action that arrives with those inputs. Today none does. We could fill the gap with an estimated clicks table per category, but that would make the order look measured while being a guess, so we chose the simpler order instead.
Ordering is half the job. For each action, Radar can generate a copy-and-paste prompt for an AI assistant such as Claude, ChatGPT or Cursor to help draft the fix, and each fix thread names the Radar tools to run again to check it. Your team makes the change.
Separate tools versus one audit
What changes when the results are read together
- Each tool reports its own score
- Disagreements look like noise between tools
- You decide the order of fixes yourself
- All 13 results sit in one report
- Disagreements become named insights
- Actions come in a stable impact and effort order
When Should You Use Separate Tools Instead?
Use separate tools when you have one specific question. If you only need to know whether GPTBot can reach your pricing page, a single crawl check answers it. Several of Radar's checks also run as standalone tools, and the free Radar check runs the six technical tools together, one check per domain.
Use a full audit when you need to know where to start, or when the results of single checks do not add up. The seven paid tools and the cross-tool insights only come with a paid audit, from $5 for one, $12 for three or $40 for ten.
Read the transcript
Does AI describe your business accurately? Let's take a closer look with Radar. Enter your domain and email, confirm the six-digit code, then run your free check. Six technical checks give you a readiness score and a first look at your site. A paid full audit covers thirteen tools, including AI-response checks across ChatGPT, Claude, Gemini and Perplexity. Inspect the responses and citations. Here, the sample report reveals confusion with another business. Review the top priority. Generate a prompt to help draft the fix. Your team reviews and makes the change. After the change, recheck. See what still needs attention. I'm Scout, your guide to Radar.
Is Radar Built on Vector or Hive?
No. An earlier version of this post had a section claiming that Radar runs on Pixelmojo's Vector and Hive products. When we checked the code on 2 October 2026, Radar's audits ran in their own modules, separate from both. The one link is small: Radar's Search Console findings job records its activity in Hive's activity log. Vector is the AI sales agent on our own site and a client build, and it does not score Radar audits.
What Changed in This Post?
We rewrote this post on 2 October 2026 as part of a review of every post on the blog. The changes:
- the promise of a complete audit in under a minute, with every tool running at once, is removed: audits run in staged batches, and no audit duration is measured;
- two average scores, from 63 and 82 audits, that contradicted each other are removed;
- the tool list now matches Radar's 13 tools, and notes that the 13th tool arrived in June 2026;
- the description of how Radar orders fixes now matches the code;
- the claim that Radar runs on our Vector and Hive products is corrected.
What This Series Covers
This is Part 4 of The AI Visibility Stack, a six-part series on how we built Radar and what we learned running it. Part 3 covers why AI answers need their own measurement. Part 5 covers which next step fits after a free check.
One Audit, 13 Tools: Questions Readers Ask
Common questions about this topic, answered.
What is a cross-tool insight in an AI visibility audit?
A cross-tool insight is a finding that only appears when two tools are read together. For example, if a site's robots.txt rules score well but the crawl check fails, Radar flags "Robots.txt fine but crawl access poor", because the policy allows bots that something else, such as a CDN or firewall, is blocking. Neither tool shows that on its own.
Does Radar run all 13 tools at the same time?
No. Since 12 August 2026, a full audit runs its tools in staged batches. Lighter tools run three at a time, and the four provider-heavy tools (the AEO audit, Citations, Source Influence and Prompt SOV) each run in a batch of their own. Each batch starts after the previous one finishes. Results appear in the dashboard as each tool completes.
What are Radar's 13 tools?
As of October 2026, the free check runs six: Crawl Check, Robots.txt, llms.txt, AI Readiness, AEO Audit and Schema Audit. A paid audit adds seven more: Reddit monitoring, plus six that read AI answers (Citations, Citation Test, Prompt SOV, Source Influence, Hallucination Check and Brand Disambiguation). The dashboard groups them into Crawlability, Content, Citations, Community and Trust.
Why is running AI visibility tools separately a problem?
Because many problems only show up as a disagreement between two results. A robots.txt checker can pass while a crawl check fails, and a crawl check can pass while no AI answer mentions you. Run separately, each tool looks fine or looks bad in isolation. Read together, the pair points to where the problem actually is.
How does Radar decide which fix to show first?
As of October 2026, every action is ordered by its impact and effort, with ties broken the same way every time. Radar has a ranking formula that weighs traffic at stake, but none of the 13 tools measures clicks, so no action carries that input yet and the formula does not run. We would rather show a simple, stable order than an estimate that looks measured.
Can I run Radar tools one at a time?
Yes. Several of the checks also run as standalone tools at pixelmojo.io/tools, and Radar's free check runs the six technical tools together, one check per domain. The seven paid tools and the cross-tool insights come with a paid audit, from $5.
Is Radar built on Vector or Hive?
No. An earlier version of this post said Radar runs on Pixelmojo's Vector and Hive products. Radar's audits run in their own code. The one link we found is that Radar's Search Console findings job records its activity in Hive's activity log. Vector is our lead qualification agent and does not score Radar audits.
What changed in this post on 2 October 2026?
We rewrote it. The promise of a complete audit in under a minute, with every tool running at once, is removed, because audits run in staged batches and no audit duration is measured. Audit averages that contradicted each other are removed, the tool list now matches Radar's 13 tools, and the claim that Radar runs on our Vector and Hive products is corrected.
