--- title: AI Credit Limits description: >- Set monthly AI Credit limits at the organization, role, per-user, or service account level. breadcrumbs: Docs > Account Management > Billing > AI Credit Limits --- > For the complete documentation index, see [llms.txt](https://docs.datadoghq.com/llms.txt). # AI Credit Limits {% callout %} # Important note for users on the following Datadog sites: app.ddog-gov.com, us2.ddog-gov.com {% alert level="danger" %} This product is not supported for your selected [Datadog site](https://docs.datadoghq.com/getting_started/site.md). ({% placeholder "user-datadog-site-name" /%}). {% /alert %} {% /callout %} ## Overview{% #overview %} Organization admins can set monthly caps on AI Credit usage for the organization, for roles, and for individual users. A default per-user limit applies to every user, and admins can override it for individual users. Admins can also set limits for individual service accounts. AI Credits are shared across [Bits Chat](https://docs.datadoghq.com/bits_ai/bits_chat.md), [Bits Investigation](https://docs.datadoghq.com/bits_ai/bits_investigation.md), [Bits Code](https://docs.datadoghq.com/bits_ai/bits_code.md), [Bits Agent Builder](https://docs.datadoghq.com/actions/agents.md), and [Bits Database Optimization](https://docs.datadoghq.com/database_monitoring/bits_database_optimization.md). ## Permissions{% #permissions %} To view and set AI Credit limits, a user needs the [`billing_edit` permission](https://docs.datadoghq.com/account_management/rbac/permissions.md#billing-and-usage). ## Where to set limits{% #where-to-set-limits %} AI Credit limits are configured in [**Bits AI > AI Credits Management**](https://app.datadoghq.com/bits-ai/ai-credits-management). ## Types of limits{% #types-of-limits %} You can configure the following types of monthly limits: | Limit type | Description | | ------------------------ | ------------------------------------------------------------------------------------------------------------------------------------------------ | | Organization-wide limit | Caps total AI Credit usage across the organization for the month. No organization-wide limit is set by default. | | Default per-user limit | Caps monthly AI Credit usage for each user who isn't covered by a role-based limit or an override. | | Role-based limit | Caps monthly AI Credit usage for each user with a given [role](https://docs.datadoghq.com/account_management/rbac.md). | | Per-user override | Sets a custom monthly limit for an individual user, replacing any role-based or default per-user limit for that user. | | Service account override | Sets a custom monthly limit for an individual [service account](https://docs.datadoghq.com/account_management/org_settings/service_accounts.md). | ## How limits are applied{% #how-limits-are-applied %} When more than one limit applies, Datadog checks them in the following order: | Check | Users | Service accounts | | -------------- | ------------------------------------------------------------------------ | --------------------------- | | Always applied | Organization limit | Organization limit | | Priority order | 1. Per-user override2. Highest role-based limit3. Default per-user limit | 1. Service account override | Within the organization limit, Datadog applies the first limit in the priority order that's set and skips the rest. - Raising a limit unblocks users who reached the previous limit. Lowering a limit blocks users whose usage already exceeds the new value. - When a user or the organization reaches a limit, affected users can't use Bits Chat, Bits Investigation, Bits Code, Bits Agent Builder, or Bits Database Optimization. A banner shows the reset date. - When a limit is reached, requests to start an investigation through the [Bits AI API](https://docs.datadoghq.com/api/latest/bits-ai.md) return HTTP `429` with the error code `AI_CREDITS_EXHAUSTED`. A `429` with the error code `INVESTIGATION_RATE_LIMIT_EXCEEDED` means an [investigation rate limit](https://docs.datadoghq.com/bits_ai/bits_investigation/configure.md#configure-rate-limits) was reached instead. ### Service accounts{% #service-accounts %} Datadog reports [service account](https://docs.datadoghq.com/account_management/org_settings/service_accounts.md) usage separately from user usage. Service accounts have no default limit. Role-based limits and the default per-user limit don't apply to service accounts, even if a service account has a role with a limit. ### Examples{% #examples %} #### How the organization limit caps total usage{% #how-the-organization-limit-caps-total-usage %} These examples show how the organization limit and the per-user limit combine to cap total billable usage in a 20-user organization. In the second row, the per-user limits add up to more than the organization limit, so the organization limit becomes the effective cap. | Organization limit | Per-user limit | Users | Maximum hypothetical usage | Maximum billable AI Credits | Result | | ------------------ | ----------------------- | ----- | ----------------------------- | --------------------------- | -------------------------------------------------------------------------------------------------------- | | 2,000 AI Credits | 50 AI Credits per user | 20 | 50 × 20 = 1,000 AI Credits | 1,000 AI Credits | Each user can use up to 50 AI Credits without other restrictions. | | 2,000 AI Credits | 200 AI Credits per user | 20 | 200 × 20 = 4,000 AI Credits\* | 2,000 AI Credits | Each user can use up to 200 AI Credits until the organization-wide limit of 2,000 AI Credits is reached. | | Not set | 200 AI Credits per user | 20 | 200 × 20 = 4,000 AI Credits | 4,000 AI Credits | Each user can use up to 200 AI Credits without other restrictions. | \* A configured organization limit acts as the ceiling for the organization, regardless of the sum of hypothetical per-user limits. #### Which limit applies to a user{% #which-limit-applies-to-a-user %} In this organization, the Standard role limit is 200 AI Credits and the Admin role limit is 400 AI Credits. Each row shows a different user and the limit that applies to them, based on the priority order in the How limits are applied section. | Per-user override | User's roles | Default per-user limit | Limit applied | Why | | ----------------- | -------------------- | ---------------------- | ------------------ | ------------------------------------------------------------------------------------------ | | 50 AI Credits | Standard and Admin | 100 AI Credits | 50 AI Credits | The override is checked first, so it applies even though it's lower than both role limits. | | None | Standard and Admin | 100 AI Credits | 400 AI Credits | With no override, the highest of the user's role limits applies. | | None | Standard | 100 AI Credits | 200 AI Credits | With no override, the Standard role limit applies. | | None | No role with a limit | 100 AI Credits | 100 AI Credits | With no override and no role limit, the default per-user limit applies. | | None | No role with a limit | Not set | Organization limit | With no user-level limit configured, the user is capped only by the organization limit. | ## Usage attribution{% #usage-attribution %} Attribution determines whose limit applies to a given unit of AI usage, and who the AI Credit spend is attributed to. Usage that doesn't come from a person is attributed to **Autonomous Agents**. Within that total, usage from each service account is also broken out individually. | Product | Attribution | | ------------------------------------------------------------------------------ | ----------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- | | [Bits Chat](https://docs.datadoghq.com/bits_ai/bits_chat.md) | All Bits Chat usage, including chats inside an investigation, rolls up to the email address of the user interacting with Bits Chat. | | [Bits Investigation](https://docs.datadoghq.com/bits_ai/bits_investigation.md) | Manually started investigations roll up to the email address of the user who initiated the action.Investigations started through the API roll up to the user or service account that owns the API key, and count toward that account's limits.Automatically triggered investigations (for example, from a monitor firing) roll up to **Autonomous Agents**. | | [Bits Code](https://docs.datadoghq.com/bits_ai/bits_code.md) | All Bits Code usage rolls up to the email address of the user interacting with Bits Code. | | [Bits Agent Builder](https://docs.datadoghq.com/actions/agents.md) | Agent runs for workflows roll up to the email address of the user who created the workflow.Agent runs for workflows created by service accounts roll up to **Autonomous Agents**, and are also broken out by the individual service account. | ## Get alerted before reaching a limit{% #get-alerted-before-reaching-a-limit %} To get notified before users reach their limit, use the AI Credits [usage metrics](https://docs.datadoghq.com/account_management/billing/ai_credits.md#monitor-usage), grouped by `user_handle`, in a [metric monitor](https://docs.datadoghq.com/monitors/types/metric.md). ## Further reading{% #further-reading %} Additional helpful documentation, links, and articles: - [AI Credits](https://docs.datadoghq.com/account_management/billing/ai_credits.md)