Direct answer
Text regenerate and screenshot regenerate draw from two separate quota pools, not one shared budget — an account's text regenerate allowance and image regenerate allowance are tracked independently, alongside a separate full-generation quota for running an entire project from scratch. Understanding that these are separate pools matters because running out of one doesn't mean the other is affected, and planning regenerate usage around a single combined number leads to surprises.
Why three separate pools instead of one
Text regeneration and screenshot regeneration have different costs and different failure modes — a team might need several rounds of text regenerate to nail messaging while getting the screenshot set right on the first or second try, or the reverse. Separate pools let each kind of iteration be budgeted on its own terms instead of one type of regenerate quietly starving the other.
Full generation is the most expensive action, by design
Running a full generation — producing an entire project's output from scratch — draws from its own quota separate from either regenerate pool, and it's the heaviest action available. Teams that treat "just run it again from scratch" as a casual fix for a narrow problem burn through this quota faster than teams that diagnose the actual issue first and reach for the narrower regenerate tool that matches it.
Regenerate quota resets are plan-dependent
Free and Pro plans carry different regenerate allowances, and higher-frequency production work is one of the more common reasons teams move from free to Pro rather than working around a tighter quota. If a team is regularly running into quota limits during a normal review cadence, that's a signal to look at the plan's allowance rather than rationing reviews to fit an insufficient quota.
Common mistakes
- Assuming text and image regenerate share one pool, and being surprised when one runs out while the other still has room.
- Reaching for a full regeneration to fix a narrow problem that a single text or screenshot regenerate would have solved more cheaply.
- Not checking which quota pool is actually constrained before deciding whether an upgrade is necessary.
Checklist
- Check text and image regenerate quota separately — they don't share a budget.
- Use full generation only when starting from scratch is genuinely necessary, not as a default fix.
- If quota limits are hit regularly during normal review cycles, evaluate plan tier rather than cutting review depth.
Operating rule
Match the regenerate type to the actual problem before spending quota — the fastest way to run out of any of the three pools is using the wrong one for a problem a cheaper action would have solved.