Otter's list of the ten best meeting tools begins with Otter

Search for the best AI notetaker and you will get a ranked list. Search again and you will get twelve. They are confident, they are recent, they carry a year in the title, and almost all of them were written by one of the products on the list.

Here are four, by name, as they stood in July 2026.

Four lists

Otter, "The 10 Best AI Meeting Tools in 2026." Number one is Otter.

Humla, "The Best Local AI Meeting Notetakers for Mac (2026)." Number one is Humla.

Humla, "6 Best Granola Alternatives in 2026." Number one is Humla.

Humla, "6 Best Otter.ai Alternatives in 2026." Number one is Humla.

Nobody is shocked. A teardown of one rival's actual pricing page or of what three free voice apps really do takes a day of reading and produces something falsifiable; a ranked list takes an afternoon and produces a search result. Only one of those is worth publishing, and it is not the one the category prints. That is the point worth sitting with: the format is so thoroughly understood as advertising that a vendor winning its own ranking does not register as a claim at all. It still occupies the search result somebody typed a real question into.

The unnumbered lists are not more honest, only harder to check. Otter's "4 Best Note-Taking Tools To Boost Your Productivity" has no ranked entries and a section headed "Transform the way you take notes with Otter.ai". Its "10 Best OneNote Alternatives for 2025" closes on "Level up your notes with Otter!" The conclusion is in the furniture rather than the numbering.

Why the genre exists

Not moral failure. Incentives.

A best-of list ranks for a query with buying intent, which is the most valuable kind there is. It is cheap to produce and cheaper to reproduce: the same skeleton refills for "best for Zoom", "best for recruiters", "best free", "best for Mac", indefinitely. Otter has been running the genre since at least 2018, with versions titled for 2018, 2022, 2025 and 2026 — the same page rewritten as the calendar turns.

And it works, which is the part that matters to you. A company with a content team and a search budget can occupy most of the first page of results for a question about its own competition. The reader has no way to see that the twelve independent-looking answers are six companies.

The same incentive shapes what does not get written. Nobody publishes a page showing every file the product leaves on your disk, because that page cannot be reused for "best for recruiters" next quarter.

We do this too

We publish a comparison of nine local-first transcription tools. It has a table. yapyap is in the table. It would be absurd to run this teardown without saying so.

What we tried to do differently is narrow and checkable, so hold us to these rather than to the intention:

It names where you should buy something else. Three of the nine get an explicit "buy this instead, and here is when". MacWhisper is cheaper than us and we say so with no argument attached. Meetily is the closest real rival and we tell you to go and read its source.

yapyap is not first. The table is not ranked at all, and ours sits at the bottom of it. Our comparison pages work the same way, including the one against Otter, and so does the data-safety piece that puts three products side by side on one question.

Everything is dated, and gaps are marked as gaps. Where our research recorded nothing, the cell says "not stated" rather than an estimate. Pricing in this category moves constantly, and a comparison table without a date on it is a claim about a month that has already ended.

None of that makes it independent. We make one of the products. The editorial standards page says so at the top, and this post says it again: a comparison written by a vendor is a sales document with a method section, and the method section is the only thing that distinguishes one from another.

Three questions, twenty seconds each

Who published it? Scroll to the footer. If the site sells one of the entries, you are reading an advert with a table in it. That does not make it worthless — vendors often know the category best — but it changes what the ranking is evidence of.

Does it recommend against itself anywhere? A list where the publisher is best at everything for everyone tells you nothing. A list that says "if you need X, buy the other one" has at least been written by somebody willing to lose the sale.

Is it dated, and does it say where the numbers came from? Prices here moved materially over the past year. A table with a year in the title and no date in the body is decoration.

Apply all three to this post. Published by yapyap, which sells a competing product. The four lists above were read in July 2026 and any of them may have been rewritten since — being rewritten is the genre's defining habit. Corrections go to the corrections page, where they always have.