Anthropic Introduces Routines for Claude Code Automation

Anthropic has introduced a new feature called Routines for Claude Code, allowing developers to configure automated coding workflows that run on schedules, through API calls, or in response to external events. The feature runs on Claude Code’s cloud infrastructure, removing the need for developers to maintain their own cron jobs, servers, or automation pipelines locally. A routine consists of a prompt, repository access, and connected tools or services. Once configured, the routine can execute repeatedly without manual intervention. Scheduled routines support recurring jobs such as triaging bugs, scanning documentation drift, or generating pull requests. API-triggered routines expose endpoints and authentication tokens, enabling external systems such as deployment pipelines, monitoring platforms, or internal tooling to trigger Claude Code sessions through HTTP requests. Anthropic also introduced webhook-based routines for GitHub events. Developers can configure routines to automatically launch sessions when pull requests match specific conditions. Claude Code can then monitor updates to the pull request, respond to comments, track CI failures, and continue operating across the lifecycle of the change. The company says teams are already using routines for workflows such as automated issue triage, deployment verification, alert analysis, documentation updates, and cross-language SDK synchronization. One example described a workflow where a merged Python SDK pull request automatically triggers a routine that ports the changes into a Go SDK and opens a corresponding pull request. Another example uses monitoring alerts to trigger automated debugging and draft fixes before an engineer reviews the incident. Community reaction on X mixed enthusiasm for automation with concerns about reliability and usage limits. Developers highlighted use cases such as automated PR reviews, CI/CD workflows, and scheduled debugging tasks Lead Engineer Mike Darlington posted: Oh this is massive. AI webhooks straight from your observability dashboard or CI workflow. Others pushed back on Anthropic’s rapid release pace, calling for fixes to model degradation, downtime, and quota limits before adding more features, like user Katie Keith commenting: Ooh, this sounds good! Hopefully the new Routines can replace Schedules in Claude Cowork, which are rubbish and can only run on your own machine while you’re trying to work on something else. The release reflects a broader trend toward asynchronous AI coding agents that operate continuously in cloud environments instead of interactive local sessions. Compared to tools like Cursor, GitHub Copilot agents, and OpenAI Codex workflows, Claude Code Routines focuses more heavily on event-driven automation and persistent background execution tied directly to repositories, APIs, and operational systems.