What Is This Note?
This is a dated note on Radar's private beta, from 16 March to 2 April 2026. It replaces a post called "50 Users Ran Our Free AI Visibility Tools. Here Is What They Found." When we checked that post against our records on 2 October 2026, its central numbers did not hold up, so we cut it down to what the records show.
What Do Our Records Show?
During the beta, Radar gave access through a link sent by email. The table that stores those links is the one record we still have from the period, and it gives the numbers below. Addresses belonging to Pixelmojo are excluded.
Radar private beta, 16 March to 2 April 2026
From the access-link table, Pixelmojo addresses excluded, dates in UTC. Checked 2 October 2026
people requested access
access links sent to them
opened at least one link
tools in Radar at the time
These records show who asked for access and whether they opened the link. They do not show what anyone did after that: which tools they ran, for how long, or whether they came back.
Which Claims From the Original Post Are Withdrawn?
Every claim below appeared in the April 2026 version of this post. We checked each one against what we still hold.
| Claim in the April 2026 version | What the records show | Status |
|---|---|---|
| 50 users requested tokens | 12 people outside Pixelmojo requested access | Corrected |
| They ran all 13 tools | Radar had 12 tools until 5 June 2026; per-tool runs were not logged until 26 April 2026 | Withdrawn |
| Average session of more than 10 minutes | No session data from the period is retained | Withdrawn |
| Users came back within 48 hours to re-scan after fixes | No per-user activity records from the period | Withdrawn |
| Users whose tokens expired joined a waitlist instead of leaving | No waitlist record from the period supports it | Withdrawn |
| Cross-tool conflict detection was the feature users mentioned most | No record of user feedback ranking features | Withdrawn |
The original also drew conclusions about the market from those numbers: that the problem is real and underserved, that awareness is the bottleneck, and that cross-tool intelligence is a moat. With the numbers gone, those conclusions have nothing under them here, so we removed them too. Where we still hold a view, we argue it from evidence in other posts, such as why AI answers need their own measurement.
What Happened After the Beta?
Radar launched publicly on 19 April 2026, with a free first check and paid audits. The beta's email links were replaced by a sign-in code for the free check and a sign-in link for paid accounts.
From private beta to today
Dates from the Pixelmojo git history and launch posts
Private beta
access by email link
Public launch
free check and paid audits
13th tool
Brand Disambiguation
Staged batches
provider-heavy tools run alone
As of October 2026, the free check runs 6 technical tools on one domain, once per domain, and does not ask AI engines about your brand. A paid audit runs all 13 tools, including Reddit monitoring and the tools that sample answers from ChatGPT, Claude, Gemini and Perplexity, from $5. The Radar user guide covers how it works now, and Part 2 tells the story of how it was built, from the git history.
What This Series Covers
This is Part 1 of The AI Visibility Stack, a six-part series on how we built Radar and what we learned running it. Part 1 is now a dated note; Part 2 and the parts after it were rewritten on the same day.
Radar Beta: Questions Readers Ask
Common questions about this topic, answered.
How many people used Radar in its first beta weeks?
Our records show that 12 people outside Pixelmojo requested beta access between 16 March and 2 April 2026, and 11 of them opened at least one access link. The original version of this post said 50 users. We cannot find a record that supports that number, so we withdrew it on 2 October 2026.
Did beta users spend more than 10 minutes per session?
We cannot show that. We did not keep per-session or per-tool records from the beta weeks: Radar began logging individual tool runs on 26 April 2026. The original post reported an average session above 10 minutes, and we have removed it because no data we still hold supports it.
How many tools did Radar have during the beta?
Twelve. All 12 tools ran by default from 31 March 2026, and the 13th, Brand Disambiguation, was added on 5 June 2026. The original post said beta users ran all 13 tools, which was not possible at the time.
How does Radar work today?
Radar launched publicly on 19 April 2026. As of October 2026 it starts with a free check of 6 technical tools, one per domain, which does not ask AI engines about your brand. A paid audit runs all 13 tools, including Reddit monitoring and the tools that sample answers from ChatGPT, Claude, Gemini and Perplexity, from $5.
When did Radar's private beta start?
The first beta access link was sent on 16 March 2026 (UTC). The beta ran until Radar launched publicly on 19 April 2026, and this note covers its first two and a half weeks, to 2 April 2026.
How did people get into the beta?
By email. A person entered an email address and received a sign-in link. Our records show 43 links sent to 12 people outside Pixelmojo by 2 April 2026, so most of them received more than one link.
Why withdraw the 50-users figure instead of correcting it?
Because nothing we still hold produces it. We could not find a list, a log or an analytics export that counts 50 people in the beta, so the honest correction was to replace it with the number our access records do support: 12 people outside Pixelmojo.
What does Radar record today?
Since 26 April 2026, Radar has logged each individual tool run, and audits saved by their owners have been stored since 30 March 2026. Those records are why later posts can count audits and domains. The beta weeks came before them.
