Release
One allowance per feature, not one pooled bucket
Usage is metered in credits against a per-action rate card, with a rolling five-hour window nested inside a weekly cap.
A single pooled usage bucket has a failure mode everyone who has used one recognises: an afternoon of one kind of work takes away a different kind of work for the rest of the month. FlowFinds meters each feature separately. Exhausting the website editor does not touch the find engine; exhausting the find engine does not silence the support agent.
Every bucket carries two nested windows. A rolling five-hour window gives back-pressure inside a working session. A weekly cap sits above it, because a five-hour window on its own would let a month of spend disappear over a weekend. Neither window alone produces sane behaviour, which is why there are two.
The rate card is per action and it is published. A full website edit runs an entire agentic session, which is why it costs fifty times a single support answer rather than the same as one. The meter is charged before the run and counts exactly what the limit enforces — there is no second, hidden accounting that decides whether you are over.
When an allowance is exhausted you are told. You are never quietly served free work, and never quietly served worse work. Pro is the reference tier: every published allowance is Pro's, and every other tier is a stated multiple of it, so there is one set of numbers on the page instead of four.
Read next
Read the rate card and the plan multiples.
More from FlowFinds
- Newer: Exclusivity is enforced by the database, not by application logic
- Older: Your first sale is funded by us
- Everything else is in the news index.
To fact-check anything above before you publish it, write to [email protected].