Anthropic released Claude Code Routines, a research-preview feature that automates repeatable work by packaging a Claude Code configuration into a reusable, trigger-driven execution unit. Routines run on Anthropic-managed cloud infrastructure, meaning they continue executing even when your laptop is closed.
A routine consists of a saved prompt, one or more repositories, and a set of connectors. Once created, routines can be invoked through three trigger mechanisms. Scheduled triggers run on a recurring cadence—hourly, nightly, or weekly. API triggers fire on demand by sending an HTTP POST to a per-routine endpoint with a bearer token. GitHub triggers react automatically to repository events such as pull requests, pushes, issues, or workflow runs. A single routine can combine multiple triggers simultaneously; for example, a PR review routine could run nightly, trigger from a deploy script, and also react to every new PR.
Routines are available on Pro, Max, Team, and Enterprise plans with Claude Code on the web enabled. Users can create and manage them at claude.ai/code/routines or from the CLI with the /schedule command.
The documentation outlines several concrete use cases. Backlog maintenance routines run on a schedule against an issue tracker, read issues opened since the last run, apply labels and assign owners based on code area references, and post summaries to Slack. Alert triage routines receive API calls from monitoring tools when error thresholds are crossed, correlate stack traces with recent commits, and open draft pull requests with proposed fixes. Code review routines trigger on pull requests, apply team checklists, leave inline comments for security and style issues, and add summary comments for human reviewers. Deploy verification routines run after production deploys, execute smoke checks, scan error logs for regressions, and post status updates. Docs drift routines scan merged PRs weekly for API documentation changes and open update PRs for editors to review. Library port routines use GitHub triggers on merged PRs to port SDK changes across languages automatically.
Routines execute as autonomous Claude Code cloud sessions with no permission-mode picker and no approval prompts during a run. The session can execute shell commands, use skills committed to the repository, and call included connectors. Access is scoped by repository selection, branch-push settings, environment network access variables, and included connectors. Routines belong to individual claude.ai accounts, are not shared with teammates, and count against the account's daily run allowance. Commits and pull requests created by routines appear under the routine creator's GitHub identity.
As of the documentation date, Routines remain in research preview, with behavior, limits, and the API surface subject to change.