How do you measure ChatGPT plugin visibility?
Measure three layers and keep them apart: the directory, tool selection once the plugin is installed, and suggestions to people who have not installed it. Each layer needs repeated runs on the real product and is reported as a count, such as 7 of 10 runs, never as a rank. The directory can be read by anyone without an account. Tool selection has a method documented by OpenAI. Suggestions before install are the hard part, and KailxLabs has not yet shown they can be counted reliably.
Three layers to measure
Each point carries the page it came from. Platform behaviour is described in OpenAI's own words where we can.
Search the directory for your category words and your name, and record which shelves you sit on. The directory front page showed 16 shelves and 96 slots when we read it logged out on 9 Oct 2026.
OpenAI, ChatGPT plugin directory, read logged out · KailxLabs, KailxLabs directory pilot, 9 queries and one shelf snapshot
With the plugin installed, replay a fixed set of direct, indirect and negative prompts and record which tool ran. This is the method in OpenAI's own metadata guide.
OpenAI, Developer docs, optimize metadata
On an account with nothing installed, replay the prompts in new chats and record whether any plugin is suggested, and which. State the account, plan, region and date, because OpenAI says what a user sees depends on the account.
OpenAI, Help Center, Plugins in ChatGPT
We found no documented dashboard for impressions, suggestions or installs. Your own server logs show calls to your plugin, which covers layer 2 only.
OpenAI, Developer docs, optimize metadata
What we do not know
- KailxLabs has piloted layer 1 only. Layers 2 and 3 are designed and not yet run.
- Suggestions may be rare. If they are, layer 3 will mostly report that nothing was suggested, and that is a result, not a failure.
- Results from one account cannot be pooled with another plan, region or app.
Nobody outside OpenAI knows the full recommendation method. We report what is documented, what we counted, and what is still unknown, and we keep the three apart.
Sources and dates
Source date is the date on the source itself. Checked is the day we last read it.
| Source | Publisher | Source date | Checked |
|---|---|---|---|
| ChatGPT plugin directory, read logged out | OpenAI | Live page | 9 Oct 2026 |
| KailxLabs directory pilot, 9 queries and one shelf snapshot | KailxLabs | 9 Oct 2026 | 9 Oct 2026 |
| Developer docs, optimize metadata | OpenAI | Not dated | 9 Oct 2026 |
| Help Center, Plugins in ChatGPT | OpenAI | Updated early Oct 2026 | 9 Oct 2026 |
Related: How KailxLabs measures AI answers · Can ChatGPT recommend plugins automatically? · ChatGPT plugins research hub
- 9 Oct 2026: answer published, from OpenAI pages read at the source that day.
Talk to the founder.
20 minutes with Kailesk Khumar. You leave knowing where AI names you today, and what it would take to become the answer.
- A live look. We ask five of your buyer questions on ChatGPT, Google and Perplexity while you watch.
- The reasons. The three biggest reasons AI names a rival instead of you.
- The plan. A clear next step and its price, whether or not you hire us.
Pick a time that suits you. The calendar invite arrives straight away. Prefer a link? Book on Cal.com.