Of the 46 tools we track in detail, **42 have a free tier** — a permanent one, not a trial. That is 91%, and it is high enough to be a structural fact about the market rather than a marketing choice.
Why free tiers dominate
**The switching cost is near zero**, so acquisition has to be too. A tool that asks for a card before showing value is asking a user to commit to a category they have not verified exists.
**The marginal cost of a free user is low** until they hit a limit, and the limit is where the business model lives: tokens, messages, exports, seats.
**Competition is horizontal.** When five assistants do the same thing, the free tier is the demo, and the demo is the product decision.
The four that do not
The exceptions cluster around tools where the free tier would be indistinguishable from the paid product: heavy compute, seats tied to an organisation, or a workflow that only makes sense at team scale.
That is the honest reason. A free tier for those is not generosity, it is a subsidised product that would cannibalise itself.
What to watch instead of the free tier
**The limit, not the existence.** 'Free tier available' is on every page. The question is what runs out first: requests, storage, seats, or export. Each tells you what the maker is optimising for.
**What happens at the limit.** A tool that degrades gracefully keeps you; one that locks your data punishes you for staying. Check before you put real work in.
**Whether the free tier is permanent.** A trial labelled as a free tier is a pricing page in disguise, and the switch is usually announced by email you will not read.
How we record it
Every tool entry carries free-tier status and the entry price, checked weekly and dated. When a free tier disappears, the comparison pages that depend on it get updated in the same pass — a stale free tier claim is the fastest way to lose a reader's trust.