FIG. 05
ChatGPT recommended an equity crowdfunding platform in 5 of 5 runs. The platform shut down in 2023.
stale answers persist for years
AI SEARCH VISIBILITY
Your next customer just asked ChatGPT what to buy.
nthrun checks whether the answer names you. It asks the question your buyers ask, many times, across ChatGPT, Perplexity, Gemini and Claude, and reports how often you come up with the full run-to-run spread.
25 free credits on signup. Email only, no card.
WHAT WE RAN
ChatGPT (OpenAI)
nthrun 0/10
Brand24
10
BuzzSumo
10
Mention
10
Google Alerts
9
SEMrush
8
Perplexity
nthrun 0/10
Frase
9
Brand24
6
ChatGPT
6
Otterly.AI
6
Google AI Overviews
5
Gemini (Google)
nthrun 0/10
Brand24
9
F5Bot
7
Syften
6
Google Alerts
5
Brandwatch
3
Claude (Anthropic)
nthrun 0/10
Google Alerts
10
Brand24
9
Semrush
9
Mention
8
Brandwatch
6
nthrun's own category. The four engines named these tools. They mostly disagree, and nthrun is in none of them yet.
01 · THE PROBLEM
Ask the same question five times. Get five different answers.
AI engines are non-deterministic. Independent research puts the odds of ChatGPT returning the same brand list twice in a row below 1 in 100. A tool that runs your prompt once and hands you a score is reporting a coin flip. If you show up in 4 of 10 runs, the honest answer is a range, and the range is what nthrun reports.
02 · FIELD NOTES
Things we found by running the same question on repeat.
FIG. 06
Asked for note-taking tools, Gemini ranked NotebookLM first. NotebookLM is a Google product.
engines favor their own
FIG. 07
One product scored 0% on its category phrasing and 100% when the question was reworded.
the prompt is half the measurement
Every finding above came out of a standard check. Transcripts included.
03 · HOW IT WORKS
How it works
-
01
Add your brand and prompts. The questions your customers actually ask AI. nthrun suggests prompts if you're not sure where to start.
-
02
Each prompt runs multiple times per engine. Same question, repeated, on the engines you pick.
-
03
You get mention rates with their spread. Every number ships with its run count, range, and full transcripts.
04 · METHODOLOGY
The methodology is public. That's the product.
Every check shows raw run counts. Every transcript is inspectable. Variability is reported, never smoothed over. We don't pin temperature or seed parameters, because your customers don't either.
05 · PRICING
Pay for checks. Skip the subscription.
Taster
$5
150 credits
Starter
$15
500 credits
Most popular
Growth
$40
1,500 credits
Pro
$100
4,200 credits
1 credit = 1 prompt × 1 engine × 1 run. Perplexity runs cost 3 credits; its search API costs more and we pass that through. Credits never expire. Check weekly, or once before a launch and again in six months.
06 · WHO'S BEHIND THIS
nthrun is one person. I built it because the established trackers start around $200 a month, lock you into a subscription, and hide their sampling. The first brand I tracked was nthrun itself; the chart at the top of this page is that data. If something breaks, the person answering your email wrote the code.
— Filip, founder
07 · FAQ
Frequently asked questions
What is nthrun?
nthrun is an AI-search visibility tracker. It measures how often a brand is mentioned when AI engines (ChatGPT, Perplexity, Gemini and Claude) answer the kinds of questions customers ask them. Instead of running each prompt once and reporting a single score, nthrun runs prompts repeatedly and reports the mention rate together with its run-to-run spread, plus the full response transcripts, so you can see exactly what the number is made of.
Why show a range instead of one score?
Because AI engines are non-deterministic: the same prompt can return different brands on different runs. A single-run score is a coin flip presented as a measurement. Reporting the mention rate with its spread across repeated runs tells you both how visible you are and how stable that visibility is, two different things a single number cannot express.
How do credits work?
One credit is one prompt, on one engine, for one run. A check that runs 5 prompts on 3 engines with 5 runs each costs 75 credits. You buy credits in packs (from $5) whenever you want, there is no subscription, and unused credits never expire. If a run fails for a technical reason, its credits are automatically refunded.
Why do Perplexity runs cost 3 credits?
Perplexity charges a flat search fee on every API call, which makes each run several times more expensive for us than runs on the other engines. The 3-credit price passes that real cost difference through instead of hiding it in a higher price for everything else.
Do credits expire?
No. Credits never expire. Check your brand weekly, or once before a launch and then again in six months. The credits you bought stay valid. This is deliberate: nthrun is built for solo founders and consultants who need measurements around real events, not another monthly subscription to feed.
Which AI engines are covered?
ChatGPT (OpenAI), Perplexity, Gemini (Google) and Claude (Anthropic). Each engine is queried through its standard API without pinning temperature or seed parameters, so responses vary run to run exactly as they do for real users. That variability is preserved on purpose, because it is part of the measurement.
Can't I just ask ChatGPT myself?
You can, and you should. Ask twice and you'll likely get two different lists. One run tells you what one customer saw once. nthrun repeats the question enough times to give you a rate, on four engines, with the transcripts to back it.
I've never used a visibility tool. Where do I start?
With one question a buyer would type, like "best [your category] for [your customer]". Your 25 free signup credits cover a first check on it. The result tells you whether AI already knows you exist, and the transcripts show who it names instead.
See what AI says about your brand.
25 free credits on signup. Email only, no card.