Credits dropped without an obvious agent
Normal chat replies can use credits. A visible agent block is not required for AI usage. Long prompts, attached files, large tab context, and advanced models can increase cost.A task stopped before finishing
Check the chat for a request for input, a credit message, or an error. If the companion returned a sample and you want the full batch, review the sample and ask it to continue. A small result alone does not tell you whether a plan limit was the cause.Rate-limit window is active
A rate-limit window is a short-term usage guardrail. It can appear even if monthly credits remain. Open the chat content panel to see the active window and refill time.Top-ups not used yet
Available included credits normally pay before purchased top-ups. Top-ups can also pay when a rate limit temporarily holds back monthly credits. Purchased top-ups do not expire at month end. Check Available now and the credit breakdown atstrawberry://settings/account.
Check which account pays
Openstrawberry://settings/account to see your paying account and credit breakdown. If you belong to several teams, use Paying account to choose one. On the web, use Billing & usage.
The selected team pays first. Eligible Personal plan and extra credits may be used automatically afterward; the free Personal allowance is excluded. Viewing another team in team settings does not switch who pays, and switching teams does not transfer credits.
Your team has no available credits
Checkstrawberry://settings/account for available credits and refill times. Team admins and owners can buy shared top-ups or review seat plans at strawberry://settings/team.
There is no automatic pay-as-you-go overage to enable. If no eligible team or Personal reserve credits are available, wait for a refill or ask an admin or owner for help.
Reduce usage
- Ask for a plan before execution.
- Ask for a sample before a large batch.
- Limit rows, pages, sources, or candidates.
- Use files and connected apps instead of pasting huge text.
- Stop or steer off-track chats.
- Save stable context in memory.
- Save repeated process in skills.