Claude Code Launches Cloud Scheduled Tasks: Autonomous AI Coding on Anthropic Infrastructure

Claude Code Now Runs Autonomous Scheduled Tasks in the Cloud

Anthropic has launched cloud scheduled tasks for Claude Code, allowing developers to set up autonomous AI coding workflows that run on Anthropic's infrastructure — even when the developer's computer is off. The feature quickly hit the HN front page with 111 points and 75 comments.

What Are Cloud Scheduled Tasks?

Scheduled tasks run a prompt on a recurring cadence using Anthropic-managed infrastructure. They keep working autonomously:

• PR reviews each morning
• CI failure analysis overnight with summary reports
• Documentation sync after PRs merge
• Dependency audits every week
• Code quality checks on custom schedules

Three Scheduling Options

How It Works

1. Create a task from the web, desktop app, or CLI with
2. Write a self-contained prompt describing what Claude should do
3. Select one or more GitHub repositories
4. Claude clones the repo, executes the task, and creates branches prefixed with
5. Results are available for review and merge

Why It Matters

This represents a significant evolution in AI-assisted development:

• Truly autonomous: Tasks run without any human intervention or machine running
• Scalable: Teams can automate dozens of recurring coding tasks
• Self-contained: Prompts must be explicit about goals and success criteria
• Branch-based: Changes go to separate branches, maintaining code review workflow

Availability

Available to all Claude Code on the web users: Pro, Max, Team, and Enterprise tiers. The feature is a strategic differentiator against GitHub Copilot and other AI coding assistants that still require active developer sessions.