Guide · Claude Code limits
What do you do when Claude Code says spend limit reached?
A spend limit is a cap on money that somebody set, so retrying never clears it. The message names whose limit it is: yours, your team's, your organization's, a channel's or a gateway's. Wait for the reset it names, or ask whoever set the limit to raise it. A usage window that runs out is a different stop, with its own fixes.
Which spend limit did Claude Code hit?
In Claude Code, the message says whose limit blocked you:
| Message | Whose limit | What lifts it |
|---|---|---|
You've hit your monthly spend limit | Your own usage credits, on Pro or Max | Raise it in Settings › Usage on claude.ai, or run /usage-credits |
You've hit your individual spend limit | A limit an admin set for you, on Team or Enterprise | An admin raises it; /usage-credits sends them the request |
You've hit your org's monthly spend limit | The whole organization's usage credits | Whoever manages billing, in Organization settings › Usage |
You've hit your team's shared budget | A pooled budget for a group you belong to; the message does not name the group | An admin raises it |
You've hit your channel's monthly spend limit | The budget of the one Slack channel the session runs in | An org owner or the channel's manager |
spend limit reached (daily; resets …) | A cap set by the operator of a Claude apps gateway | The reset it names, or the operator |
When a plan window is what ran out, the message also says when it resets, such as
· your session limit resets 3:45pm. Access comes back then, without anyone
raising the limit.
On an organization with usage-based billing, the message says usage limit
in place of spend limit.
Why does a Claude plan stop at a spend limit?
A plan's included usage stops covering requests when one of its windows runs out: the 5-hour session limit or a weekly limit. A request to a model that bills only to usage credits is never covered.
Usage credits pay for those requests at standard API rates. The spend limit is the cap on those credits, and the message appears once it is full.
On Pro and Max, Claude Code offers to raise or remove the limit without leaving the CLI, as long as credits remain. On Team and Enterprise, an owner sets these limits for the organization, for a group, or for one member.
The windows themselves, and when each resets, are in what to do when Claude Code says rate limit reached.
What does an API key return at a spend limit?
On a Claude Console API key, a spend limit comes back in one of three ways:
| Response | What it means | When it lifts |
|---|---|---|
400 You have reached your specified API usage limits | A limit you set for the organization, or for a workspace (…specified workspace API usage limits) | At the time the message states, or once you raise or remove the limit |
429 with enforced_spend_limit_reached and no retry-after | Your usage tier's monthly cap: $500 on Start, $1,000 on Build, $200,000 on Scale | 00:00 UTC on the first day of next month, or sooner on a higher tier |
429 with a retry-after | The limit set on the Claude Code workspace | At the retry-after time |
error.details.error_code. Retrying fails until access resumes,
including the SDK's automatic retries.
Credit balance is too low is not a limit: the Console's prepaid credits ran
out. On a Claude plan it can also mean an ANTHROPIC_API_KEY is sending your
requests through a key instead. /status shows which.
Can you set a Claude Code spend limit per developer?
On a Team or Enterprise plan with usage credits, yes: an owner can set a member's own limit. Where every developer has a seat and all extra usage goes through usage credits, that is the simplest answer.
On API keys, the Console sets spend limits for an organization or a workspace, not one person. The default workspace cannot have one. Anthropic's costs guide points to a gateway for limits per user, such as its own Claude apps gateway.
What does a Selan spend limit do?
Selan runs Claude Code and Codex on the subscriptions and API keys your company already pays for, and has two kinds of spend limit. A key's limit caps everything spent through one of the company's keys, by anyone. A person's limit is set once and applies across every key in the company.
Both run by day, week or month in UTC, and both bind the moment they are reached. The
request comes back as 402 over_budget, with a
retry-after set to the moment the period resets:
selan: you have reached your spend limit — resumes 2026-11-01 00:00 UTC
selan: this workspace has reached its spend limit — resumes 2026-11-01 00:00 UTCIf two limits are over at once, the message gives the later reset. Spend is counted per person, so one developer running hot does not use up anyone else's allowance.
A person's limit stops only what is billed per token. The company's subscription seats keep serving them after they reach it. Each developer's own budget is a bar on the statusline, so it shows before the refusal does. A CI token can have a limit of its own.
What the refusal says is in the docs on limits and the API's error table. How an owner sets a limit per person is on the spend page.
Free for up to three seats, and the Free plan needs no card.
The short version
- A spend limit is a cap someone set on paid usage; retrying never clears it.
- The message names whose limit it is: yours, your team's, your organization's, a channel's or a gateway's.
- On an API key, a limit you set returns 400, and the tier's monthly cap returns 429 with
enforced_spend_limit_reached. - Through Selan, a limit per person applies across every key, and subscription seats keep serving after it.
What Claude Code, Claude's plans and the Claude API do at a spend limit is as Anthropic's Claude Code error reference and costs guide (code.claude.com/docs/en/errors and /costs) and its API rate limits page (platform.claude.com/docs/en/api/rate-limits) stated it on 27 October 2026. It changes: check the date before relying on it. What Selan does is as its gateway, control app and docs stated it on that date.