Claude Code Just Got Routines — And It Changes Everything About How AI Works For You

Claude Code Routines are saved AI automations that run on Anthropic’s cloud infrastructure on a schedule, via API trigger, or in response to GitHub events — without the user’s device needing to be active.

Until now, AI coding assistants had one big limitation: they needed you to be present. You open the tool, type your prompt, watch it work — and the moment you close your laptop, everything stops.

That just changed. On April 14, 2026, Anthropic released a new feature called Routines for Claude Code. It’s currently in research preview, but the implications are massive — and you don’t need to be a developer to understand why this matters.

Here’s the beginner-friendly breakdown.

What Is Claude Code, First of All?

Claude Code is Anthropic’s AI-powered coding tool — part of the Claude app. Think of it as having an expert software engineer sitting beside you, available any time. You describe what you want in plain language, and Claude Code writes, fixes, or improves code for you.

Developers use it to build apps, fix bugs, automate repetitive tasks, and review code — all through a conversational interface.

So What Are “Routines”?

A Routine is a saved set of instructions that Claude Code can run on its own — without you being logged in or your laptop being open.

Think of it like setting up a task for a very capable assistant: “Every morning at 6 AM, check if there are new bugs in my app, write a fix, and send me the report.” You set it once. It runs automatically, on Anthropic’s cloud servers.

Simple definition: A Routine = a prompt + a codebase + connected tools, packaged once and run automatically on a schedule or trigger.

What Can Trigger a Routine?

Routines can be set off in three different ways:

1. On a Schedule (like an alarm clock)

You tell Claude Code: “Run this every night at 2 AM” or “Run this every Monday morning.” It simply does it, on time, every time — even if your device is off.

2. Via an API Call (like a remote button)

Developers can trigger a Routine by sending a signal from another system. For example, if a server somewhere detects a problem, it can automatically ping Claude Code to investigate and fix it.

3. From GitHub Events (like an automatic reaction)

GitHub is the platform where most software code lives. If a developer submits a new piece of code (called a “pull request”), Claude Code can automatically review it, check for issues, and post a summary — all without anyone asking it to.

Real-World Examples — What Can Routines Actually Do?

How Is This Different from Cron Jobs or Github Actions?

You might have heard of things like “cron jobs” or “GitHub Actions” — these are existing ways to automate tasks on a schedule. Routines are different in one key way:

The difference: Traditional automations run fixed scripts — the same steps every time, no matter what. Claude Code Routines are AI-driven — Claude reads the situation, understands the context, and decides what to do. It’s the difference between a timer and an intelligent assistant.

We will cover an in-depth comparison in another article. Stay tuned 🙂

Who Can Use This? (Plan & Limits)

Routines are available to Claude Code users on the following plans:

• Pro — 5 routine runs per day
• Max — 15 routine runs per day
• Team — 25 routine runs per day
• Enterprise — 25 routine runs per day

Additional runs beyond the daily cap draw from your standard usage allocation. And since this is a research preview, these limits may change as Anthropic refines the feature.

Why Does This Matter — Even If You’re Not a Developer?

The bigger picture here is what Routines represent: AI that works for you in the background, not just when you’re actively using it.

We’re moving from AI as a tool you pick up and put down — to AI as something you configure once and it keeps running, quietly and independently, handling work while you sleep.

For businesses, this means fewer manual processes. For developers, this means hours reclaimed every week. And for the rest of us, it’s a preview of a world where software takes care of itself.

Bottom Line

Claude Code’s new Routines feature is a genuine leap forward. It turns Claude from a reactive assistant into a proactive, autonomous worker — one that runs on Anthropic’s cloud, needs no babysitting, and can respond to real-world events in real time.

It’s currently in research preview, so expect it to evolve. But the direction is crystal clear: AI that doesn’t wait for you — it gets the job done while you’re away.

Frequently Asked Questions (FAQ)

Q: What are Claude Code Routines?

Claude Code Routines are saved AI automations that combine a prompt, one or more code repositories, and connected tools — packaged once and run automatically. They execute on Anthropic’s cloud servers on a schedule, from an API call, or in response to a GitHub event, with no need for your laptop to be switched on.

Q: Is Claude Code Routines free to use?

Routines are included with paid Claude plans — Pro, Max, Team, and Enterprise. There is no separate free tier. Pro users get 5 routine runs per day, Max gets 15, and Team/Enterprise gets 25. If you exceed the daily cap, additional runs draw from your subscription’s extra usage allocation.

Q: Do I need to be a developer to use Claude Code Routines?

Routines are built for developers and technical teams working with code repositories. Setting one up requires familiarity with GitHub, code repositories, and how to write a clear prompt. Non-developers can understand what Routines do, but actually creating and managing them requires some technical background.

Q: How is Claude Code Routines different from GitHub Actions or cron jobs?

GitHub Actions and cron jobs run fixed, pre-written scripts — the same steps every time, regardless of what they find. Claude Code Routines use an AI model to read the situation and decide how to respond based on context. For mechanical, repeatable tasks like building and testing, GitHub Actions is faster. For tasks that need judgment — like reviewing code for security issues or triaging production alerts — Routines are more powerful because they adapt to what they encounter.

Q: What can trigger a Claude Code Routine?

A Routine supports three trigger types: (1) Schedule — runs on a recurring cadence like hourly, nightly, or weekly; (2) API call — any external system can send an HTTP POST request to a per-routine endpoint to trigger it on demand; (3) GitHub event — such as a pull request being opened, a push, a new issue, or a workflow run. A single routine can combine all three trigger types at the same time.

Q: Does a Claude Code Routine keep running when my laptop is off?

Yes. Routines run entirely on Anthropic’s managed cloud infrastructure. Your device does not need to be online. The routine executes, logs its session at claude.ai/code, and can open pull requests, post to Slack, or perform other connected actions — all independently.

Q: Is Claude Code Routines fully launched, or still experimental?

As of April 14, 2026, Claude Code Routines is in research preview. The feature works and is available on Pro, Max, Team, and Enterprise plans, but Anthropic may change limits, API behaviour, and functionality while gathering feedback. The API runs under a dated beta header (experimental-cc-routine-2026-04-01), which signals it may evolve.

Q: Where do I set up a Claude Code Routine?

You can create and manage Routines at claude.ai/code/routines in your browser. From the command line, use the /schedule command inside Claude Code. You need Claude Code with web access enabled on your account. Any prior /schedule tasks you created migrate automatically — no manual steps needed.

Q: Can one Routine have multiple triggers?

Yes. A single Routine can combine a schedule trigger, an API trigger, and a GitHub trigger — all active at the same time. For example, a code review routine can run nightly, also be triggered from your deployment pipeline via API, and react to every new pull request from GitHub.

Q: What tools and services can a Routine connect to?

Routines ship with access to your connected repositories and any connectors you have configured — such as Slack, Linear, Google Drive, and GitHub. Each execution is a full Claude Code session, meaning it can run shell commands, use skills committed to the cloned repository, and call any MCP connector you have set up.