On the Claude Team plan, you can't set one limit for a team: set a monthly spend limit on each member's usage credits, plus one for the organization, in Organization settings > Usage. Seat-based Enterprise adds limits by seat tier and group. On API billing, the Claude Code workspace in the Claude Console supports a monthly limit per user and per workspace. Limits block usage until the month ends. Measure each team's spend first, then set member limits that add up to the team's budget.
What Claude Code spend can you limit, by billing type?
| How you pay | What you can limit | Where | Who sets it |
|---|---|---|---|
| Claude Team seats | Usage credits: organization monthly limit, monthly limit per member | claude.ai: Organization settings > Usage | Owners and Primary Owners |
| Seat-based Claude Enterprise | Usage credits: organization, seat tier (Standard or Premium), group and member | claude.ai: Organization settings > Usage | Owners and Primary Owners |
| API billing (Claude Console) | Monthly limit per user and per workspace in the Claude Code workspace; rate limits; your organization's monthly cap | Console: Settings > Workspaces and Settings > Billing | Organization admins, who administer every workspace |
| Amazon Bedrock, Google Vertex AI, Microsoft Foundry | Your cloud provider's budget controls | Your cloud billing console | Your cloud admins |
On seat plans, the seat's own allowance is the first ceiling: each member's usage resets on a rolling five-hour window and a weekly window, shared with Claude chat and Cowork. Spend limits only apply to the usage credits that let members keep working past it. Source: Anthropic's Claude Code cost guide.
How do spend limits work on Claude Team and Enterprise seats?
- 01Turn on usage credits
An Owner opens Organization settings > Usage on claude.ai and enables usage credits. Team plans pay for credits up front, with optional auto-reload; seat-based Enterprise plans are billed at the end of each month for actual use. Credits are billed at standard API rates.
- 02Set the organization limit
A monthly limit on all usage credits in the organization. Changes take effect immediately.
- 03Set member limits
Under Spend limits by user, give each member a monthly limit. On seat-based Enterprise, you can also set limits for everyone on a seat tier, or by group.
- 04Choose who gets credits
Enable usage credits for specific people or the whole organization. Leaving someone off keeps them on their seat allowance only.
Team-plan credits expire 12 months after purchase and are non-refundable, so don't pre-buy far more than your teams use. Usage credits don't apply to usage-based Enterprise plans. Details: Manage usage credits for Team and seat-based Enterprise plans.
What happens when a developer hits a Claude Code limit?
- Seat allowance: Claude Code says the developer hit their session or weekly limit and shows when the window resets. With usage credits on, they can keep going on credits.
- Member spend limit: the developer's usage credits pause until the end of the month. Claude Code names the limit they hit (individual, organization or team budget).
- Organization spend limit: nobody in the organization can use Claude, Cowork or Claude Code on credits until the next billing period, or until an Owner raises the limit.
- API workspace limit: requests fail with an error that starts "You have reached your specified workspace API usage limits" and says when access resumes.
Limits block rather than warn, and a request already in flight can take spend slightly past the limit. Developers can run /usage to see their own spend this month against the limit that applies to them, and /usage-credits to send their admins a request for more (on seat-based Enterprise, admins get a daily email of open requests). /usage-credits isn't available when Claude Code signs in with an API key.
How do I limit Claude Code spend on API billing?
When a developer first signs in to Claude Code with a Claude Console account, Anthropic creates a Claude Code workspace in your organization and issues that developer their own API key in it. It is the only Console workspace that supports monthly spend limits per user.
- Workspace limit: in Settings > Workspaces, select the Claude Code workspace and set a monthly spend limit, with alerts when spend reaches thresholds you choose. It can be lower than, but not higher than, your organization's limit.
- Rate limits: cap requests and tokens per minute for the workspace per model, so one runaway session can't use your whole organization's rate limit.
- Organization cap: your usage tier has a monthly spend cap ($500 on Start, $1,000 on Build, $200,000 on Scale), and you can set a lower one in Billing.
Don't archive the Claude Code workspace to stop spend: that disables Claude Code sign-in through Console billing for the whole organization. Sources: Claude Console workspaces and rate limits.
How do I limit Claude Code spend on Bedrock, Vertex AI or Foundry?
Claude Code on Amazon Bedrock, Google Cloud's Agent Platform (Vertex AI) or Microsoft Foundry is billed per token to your cloud account, so spend controls are your cloud provider's budgets. Anthropic's analytics don't cover this usage. For limits per developer on these providers, Anthropic's cost guide points to its self-hosted Claude apps gateway, which adds per-user spend limits.
How do I turn member limits into a budget for each team?
Without a team-level control, a team's budget is the sum of its members' limits. Set those limits from what the team actually spends. An illustrative team on API billing, with each developer's monthly Claude Code cost measured at API rates:
| Measure | Amount | What to do with it |
|---|---|---|
| Median developer | $190/month | Most developers sit here |
| 90th percentile | $420/month | Set the default member limit just above: $450 |
| Heaviest developer | $610/month | Give the 2 heaviest users their own limit: $700 |
| Team ceiling | $5,000/month | 8 × $450 + 2 × $700; compare with the team's budget |
On Team seats, apply the same idea to usage credits only: most developers may need a small credit limit or none, while the few who hit their weekly limit get a larger one, or a Premium seat. Set the organization limit to the sum of all teams' ceilings plus a buffer, and review limits monthly. If a team's ceiling is above its budget, the fix is usually the default model, not tighter limits: see Opus or Sonnet in Claude Code.
What do Anthropic's spend reports leave out?
Team and Enterprise owners can export a spend report from Analytics on claude.ai: one row per person per model, with requests, tokens and spend, refreshed daily with a one-day delay. Claude Code analytics adds cost per session, commit and pull request, and top spenders. Three things are missing for team budgets:
- Your teams. Reports are per user. Grouping by team or cost center is a spreadsheet job every month.
- Usage inside the seat allowance. It isn't metered in dollars, so on seat plans the report shows usage credits only, and a developer who never passes the allowance looks free.
- Cloud provider usage. Bedrock, Vertex AI and Foundry usage isn't in Anthropic's analytics.
How do I see Claude Code spend by team before setting limits?
Send Claude Code's OpenTelemetry export to Ganivra, with a team.id on each machine. The Coding agents tab then shows spend per team and per developer at API rates, on seats, API billing and cloud providers alike, including usage inside seat allowances. That gives you the median, 90th percentile and heaviest developer for every team in the table above.
- Roll it out once to every machine with managed settings: see Claude Code managed settings for cost tracking.
- Set a Ganivra budget for each team or developer, matching the same
team.idor developer as the Coding agents tab, or on all Claude Code spend or one model such as Opus. Warning and critical thresholds show in the dashboard as spend crosses them, so a team lead sees a team heading for its ceiling before Anthropic blocks anyone. Ganivra alerts; the limits that block usage stay in Anthropic's console. - Review each month which teams and developers are close to their limits, and whether seat tiers or API billing would cost less: see Claude Code Team seats vs API.
Claude Code spend limits: frequently asked questions
Can I set a Claude Code spend limit per team?
Not as a single team limit on the Claude Team plan: it offers an organization-wide monthly limit and a monthly limit per member. Seat-based Enterprise adds limits by seat tier and by group. On API billing, the Claude Code workspace supports a monthly limit per user and one for the workspace. To cap a team on the Team plan or the API, set each member's limit so the team's limits add up to its budget.
What happens when a developer hits a Claude Code spend limit?
They're blocked, not warned. On Team and Enterprise seats, Claude Code shows that the individual, organization or team budget limit was reached, and usage credits pause until the end of the month unless an owner raises the limit. On API billing, requests return an error saying the API or workspace usage limit was reached and when access resumes.
What are Claude usage credits?
Usage credits let Team and seat-based Enterprise members keep using Claude and Claude Code after their seat allowance runs out. They are billed at standard API rates. Team plans buy them in advance, and they expire 12 months after purchase and are non-refundable; seat-based Enterprise plans are billed monthly for actual use.
Who can set Claude Code spend limits?
On Team and seat-based Enterprise plans, Owners and Primary Owners set spend limits in Organization settings > Usage on claude.ai. On API billing, workspace spend limits are set in the Claude Console under Settings > Workspaces.
Do Anthropic's spend reports show cost per team?
They show spend per user and model, not per team or cost center. On seat plans, usage inside the seat allowance isn't metered in dollars, so the report covers usage credits only. Usage through Amazon Bedrock or Google Vertex AI isn't in Anthropic's analytics at all.
How do I decide what limit to set for each team?
Measure first. Send Claude Code's OpenTelemetry export to Ganivra for two to four weeks, label each machine with its team, and look at each team's cost per developer at API rates, including the usage inside seat allowances. Set member limits around what most developers actually use, and give the few heavy users their own higher limit.
Can I get a warning before a team reaches its Claude Code limit?
Anthropic's seat-plan limits block at the limit, and only Console workspace limits offer threshold alerts. In Ganivra, you can set a monthly budget per team or per developer with warning and critical thresholds, so the dashboard flags a team heading for its ceiling while there is still time to raise a limit or change the default model.
Before you set limits
See Claude Code spend by team and developer.
Start a 14-day Ganivra trial, turn on Claude Code's telemetry, and see what each team spends at API rates, on any plan. No credit card required.