Your org is out of extra usage. We let your admin know.

It means your team or enterprise account on Claude has exhausted the extra usage pool, the overage credit that sits on top of the included usage in your plan. The included usage refreshes on a fixed cycle (weekly on Pro and Max, monthly on Team and Enterprise). The extra usage is what the admin pre-purchased or pre-authorized so that members could keep going once the included tier was burnt through for the period. When that extra pool runs to zero, Anthropic stops new requests on the seat, surfaces this banner, and emails the workspace owner or primary admin. It is not a payment failure on your personal card. It is the org wallet being empty.

Your org's. The message says 'your org is out of extra usage' for a reason: the seat you are using is paid for and metered against the team account. Anthropic separates the included weekly or monthly bucket from the extra usage pool, and both live on the org, not on you. If you signed in to the same Claude account at home on a personal Pro plan, you would get a separate budget because that plan is billed to you and metered against you. The team seat draws from the team. There is no per-seat sub-limit you can lift from your side; only the workspace owner can authorize more spend.

Inside the Anthropic Console the admin lands on Settings, then Plans and Billing, where there is a row for Extra Usage with a current balance and a toggle for an automatic top-up cap. The fix is one of three actions: turn on extra usage (or raise the existing cap), buy a fixed amount of overage credit, or wait for the next included-usage reset to roll over. If the org is on Enterprise, the same controls are gated behind whichever finance owner provisioned the workspace, so the admin may have to escalate. The admin also gets a notification email titled along the lines of 'Your workspace has run out of extra usage' with a button back to the same settings page.

Two reasons usually. First, Claude Code, Skills, and any third-party app that signs in with your Claude account (Cursor, Zed, the various ACP wrappers) now draws from the same shared bucket as the web app, so the team's combined spend climbs faster than the dashboard makes obvious. Second, on Team and Enterprise the included usage is monthly and a single long agent run can chew through a meaningful slice of the period in one afternoon. The pool you depend on is the sum of every member's chat, every Claude Code session, and every API call routed through the workspace. One person's Claude Code run-loop can drain it for everyone.

Extra usage does not refill on its own. It is a prepaid bucket the org tops up. What refills is the included usage, and it refills on the plan's cycle: Claude Pro and Max reset every week from the day you subscribed, Team and Enterprise reset on a monthly cycle tied to the billing date. Until that next reset, you only get traffic if the admin re-enables or refills the extra usage pool. If the reset is hours away, waiting is the cheapest fix. If it is two weeks away, the admin has to act.

The first lever is the auto top-up cap in Plans and Billing. You can leave extra usage off entirely and accept that members will hit the wall at the included-usage line for the rest of the cycle. The second is to set a small monthly cap so a runaway agent cannot quietly bill four figures. The third is to move heavy users (the engineers running Claude Code all day) onto their own Pro or Max subscriptions billed to themselves, and reserve the team account for chat and lighter work. That third option is what most teams converge to after the first time this banner fires unexpectedly.

If you have a personal Claude Pro or Max subscription on a separate Anthropic account, yes. Sign that account into any tool that lets you bring your own credentials and your usage hits your personal plan instead of the empty org pool. The web chat does not let you mix accounts mid-session, but most desktop tools that wrap Claude Code do let you authenticate as yourself. Fazm is one of them: it wraps the real Claude Code agent loop and signs in with whichever Anthropic account you point it at, so a member whose org pool is dry can switch to their personal Max plan and keep going without changing tools.

Anthropic prices extra usage at standard API rates, so you are paying the per-token meter for everything past the included allotment. The included tier is the discounted bundle that comes with the plan price; extra usage is the safety valve at retail. That is why a single Claude Code session that streams a large repo and several long tool results can put real money on the meter once the included tier is gone. The dashboard shows the running extra usage spend so you can see the slope before the email lands.

Your team conversations stay on the team account. They are not portable to a personal Pro plan, and they keep working the moment the org pool refills. What you keep working on under your personal plan is whatever you start there. Most engineers running into this in Claude Code do not care; the unit of work is a session, not a chat history, and the file changes are in git. A desktop wrapper that preserves sessions on disk (Fazm does, the upstream CLI does too in ~/.claude/projects) means you can resume the same conversation under whichever account is still funded.