Anthropic launched its Compliance API on March 30 2026. Programmatic access to admin and resource audit logs for Claude Platform — built for finance, healthcare and legal. The enterprise trust signal regulated industries needed.
The Anthropic Compliance API gives organisations programmatic access to audit logs across their Claude Platform deployment. It tracks two event categories. Admin and system activities — adding or removing workspace members, creating or rotating API keys, modifying system configurations and any administrative action taken on the Claude Platform account. Resource activities — file creation, deletion and resource-level operations by Claude agents or users within the workspace. The API deliberately does not log model inference interactions — conversation content between users and Claude is not captured. This distinction matters for privacy and legal reasons: organisations can audit who did what at the system and resource level without Anthropic holding conversation records. For organisations already on Claude Enterprise, the Compliance API allows consolidation of multiple workspace audit feeds under a single parent organisation — unified visibility across all deployments.
Finance, healthcare and legal organisations face a structural compliance requirement that has blocked meaningful AI agent adoption: you cannot deploy a system that takes actions — creating files, calling APIs, modifying records — in a regulated environment without an immutable audit trail of what the system did and when. SOX, HIPAA and legal professional conduct rules all require audit trails for systems handling sensitive operations. Before the Compliance API, Claude could be used for reading and analysis tasks freely, but deploying Claude agents with write permissions required custom audit infrastructure — a significant engineering investment that most compliance teams would not approve without established vendor support. The Compliance API removes that blocker. Organisations can now prove to auditors, regulators and internal compliance functions that every admin action and resource modification by Claude is captured in a tamper-evident log accessible via API.
The Compliance API uses Anthropic's standard API authentication — your existing Claude Enterprise API key authenticates to the compliance endpoint. Event streams are paginated and filterable by event type, timestamp range, user and resource. For SIEM integration the recommended approach is polling on a scheduled interval and forwarding events to Splunk, Datadog, Elastic or equivalent. The event schema is documented in Anthropic's developer documentation and events are delivered in JSON format compatible with standard log processing pipelines. For organisations with multiple Claude workspaces, the parent organisation consolidation feature returns events across all child workspaces in a single authenticated request — eliminating the need to poll multiple endpoints separately.
Microsoft Azure OpenAI Service has offered audit logging for enterprise deployments for over a year — the Compliance API brings Anthropic to feature parity. Google Cloud Vertex AI similarly provides audit logging via Google Cloud Audit Logs. Anthropic differentiates on multi-workspace parent organisation consolidation — a capability enterprise customers with multiple business unit deployments have specifically requested. The compliance posture also signals Anthropic's intent for regulated industry expansion. The company's approaching $19 billion ARR trajectory requires significant enterprise revenue and regulated industries represent the largest untapped opportunity for AI platforms that can demonstrate appropriate compliance infrastructure. The Compliance API is both a technical feature and a strategic market signal.
Honest breakdowns every week. Zero sponsorships. Zero fluff.
Subscribe Free →