Anthropic

Anthropic is an AI safety company that builds Claude, a family of large language models. This source enables you to extract comprehensive data from the Anthropic Admin API, including Claude Code usage metrics, API usage reports, cost data, and organization management information.

To set up an Anthropic connection, you need to have an Admin API key that has the necessary permissions for the resources you want to access.

Admin API Key Required

This source requires an Admin API key which is different from standard API keys. Only organization members with the admin role can provision Admin API keys through the Anthropic Console.

The Admin API is unavailable for individual accounts. To use this source, you must have an organization set up in Console → Settings → Organization.

Set up a connection

Anthropic connections are defined using the following properties:

• name: The name to identify this connection
• api_key: Your Anthropic Admin API key (must start with sk-ant-admin...) (required)

connections.yml

connections:
  anthropic:
    - name: "my_anthropic"
      api_key: "sk-ant-admin-..."

You can also use environment variables in your connections.yml by using the {{ env_var("ENV_VAR_NAME") }} syntax.

For example:

connections:
  anthropic:
    - name: "my_anthropic"
      api_key: "{{ env_var('ANTHROPIC_API_KEY') }}"

Supported Data Assets

Anthropic assets will be ingested to your data warehouse as defined in the destination table.

Asset	Table Name	Incremental Key	Description
Claude Code Usage	claude_code_usage	date	Daily aggregated usage metrics for Claude Code users in your organization
Usage Report	usage_report	replace	Detailed token usage metrics from the Messages API, aggregated by time bucket
Cost Report	cost_report	replace	Aggregated cost data broken down by workspace and cost description
Organization	organization	replace	Information about your Anthropic organization
Workspaces	workspaces	replace	All workspaces in your organization
API Keys	api_keys	replace	All API keys in your organization
Invites	invites	replace	All pending organization invites
Users	users	replace	All users in your organization
Workspace Members	workspace_members	replace	Workspace membership information

Asset-Specific Configuration

Claude Code Usage

The claude_code_usage table supports incremental loading based on the date field. This data helps you analyze developer productivity and monitor Claude Code adoption.

Usage Report

The usage_report table contains detailed token usage metrics from the Messages API, aggregated by time bucket, workspace, API key, model, and service tier. Supports date range filtering.

Cost Report

The cost_report table contains aggregated cost data broken down by workspace and cost description. Supports date range filtering.

Organization Data

Tables like organization, workspaces, api_keys, invites, users, and workspace_members use full refresh mode as they represent current state data.

Notes

• Authentication: The Anthropic Admin API uses Bearer token authentication. Make sure your API key has the necessary permissions for the resources you want to access.
• Incremental Loading: Only claude_code_usage supports incremental loading. Other tables use full refresh (replace strategy).
• Date Filtering: usage_report and cost_report support date range filtering with --interval-start and --interval-end.
• Permissions: Some endpoints may return a 403 Forbidden error if your API key doesn't have the required permissions.
• Data Freshness: Claude Code analytics data typically appears within 1 hour of user activity completion. The API provides daily aggregated metrics only.
• Rate Limits: The Anthropic Admin API has rate limits in place. The source handles pagination automatically and respects these limits.
• Scope: This source only tracks Claude Code usage on the Anthropic API (1st party). Usage on Amazon Bedrock, Google Vertex AI, or other third-party platforms is not included.
• Timezone: All dates and timestamps are in UTC.
• Organization Access: The source requires organization-level access (not available for individual accounts).

Example pipeline

Here's an example of an asset ingesting data from Anthropic to a Snowflake table:

/* @bruin

name: anthropic.claude_code_usage
type: ingestr

@ingestr
source_connection: anthropic
source_table: claude_code_usage

destination: snowflake

@end
*/

select * from {{ source() }}

The ingestr operator will automatically pull data from the claude_code_usage endpoint of your Anthropic Admin API and load it into your target data warehouse.