Anthropic Debuts ‘Repeatable Routines’ in Major Claude Code Automation Update

Claude is stepping out of the chat box and into your workflow.

Anthropic is expanding the capabilities of its developer-focused AI tooling with a fresh update to Claude Code, introducing a new feature called “repeatable routines” alongside a redesigned desktop experience. The updates show the company’s growing ambition to streamline software development workflows and position Claude as a more central platform for coding automation.

But the bigger story is how these updates could reshape what developers actually need to do themselves.

Automating development with AI-powered routines

The update’s new routines feature allows developers to automate tasks on a schedule or trigger them in response to specific events directly within Claude Code. While automation itself is nothing new, Anthropic aims to simplify a process that has traditionally required integrating multiple tools and managing backend infrastructure.

Instead of dealing with traditional automation setups involving cron jobs, server maintenance, or configuring third-party services, developers can now package everything into a single routine. Each one includes a prompt, access to code repositories, and any necessary integrations. This enables a range of use cases, including:

• Automated API workflows
• Scheduled maintenance tasks
• GitHub-related processes

Furthermore, developers no longer need to worry about whether their local environment is running or whether a script failed silently overnight. Routines are executed on Anthropic’s own cloud infrastructure, so tasks can continue running even when a user’s local machine is offline.

The concept echoes familiar tools like cron jobs and GitHub Actions, but differs in that instead of executing static scripts, Claude Code routines leverage AI prompts. This allows them to dynamically adapt their behavior based on context and perform tasks that would traditionally require multiple layers of scripting and monitoring.

Not quite agents, but more than just scripts

Still, routines aren’t quite fully autonomous AI agents.

Anthropic is intentionally keeping routines lightweight and task-specific, operating as short-lived processes triggered by predefined conditions. Unlike agents that maintain a persistent state and continuously interact with external systems, routines are designed to execute targeted tasks and then stop.

What makes these routines interesting is how they exist between traditional automation and full-blown AI agents. Developers gain the flexibility of AI-driven decision-making without the complexity and unpredictability that can come with long-running autonomous agents. According to Anthropic, this makes routines particularly useful for tasks such as verifying software deployments, triaging alerts, and generating periodic reports.

The feature is currently being released as a research preview and is available to paid Claude Code users across Pro, Max, Team, and Enterprise tiers. Usage is capped by subscription level: Pro users can have up to 5 routines per day, Max users can have up to 15, and Team and Enterprise users can have up to 25. Additional usage is available on a pay-as-you-go basis.

Redesigning the developer experience

Alongside its routines, Anthropic has rolled out a redesigned Claude Code desktop application to reduce friction in the development process. The update consolidates several tools directly into the app, minimizing the need for developers to switch between environments.

Key additions also include an integrated terminal, in-app file editing, HTML and PDF preview capabilities, and a faster diff viewer for reviewing code changes. The layout is customizable through drag-and-drop, allowing developers to tailor their workspace according to their preferences.

Anthropic is attempting to bring more of the development workflow into a single environment, reducing reliance on external editors or plugins. By doing so, the company is positioning Claude Code as a primary interface for interacting with code.

Moving toward unified AI development platforms

If these moves indicate anything, it’s that Anthropic intends to deepen Claude Code’s role in the development lifecycle. By combining automation, AI-driven decision-making, and an integrated workspace, the company is aiming to reduce reliance on fragmented tooling and external platforms.

As competition intensifies, features like routines may become a key differentiator in how effectively AI tools can move beyond simple code generation to orchestrating complex, real-world workflows.

Want the bigger picture on Claude? Explore our full guide to Anthropic’s AI and where it’s headed.