How to Run Claude Code in the Cloud: Setup Guide (2026)

To run Claude Code in the cloud, you can use Murmur (which runs Claude Code sessions in cloud VMs with your own API keys), Cursor Background Agents (which support Claude models), or spin up your own cloud VMs with Claude Code installed. Each approach has different tradeoffs around orchestration, cost, and control. This guide covers the options and when each makes sense.

TL;DR: Claude Code Cloud Options

Pick Murmur for fleet orchestration with parallel Claude Code agents and post-PR automation. Pick Cursor Background Agents if you already use Cursor and want cloud tasks without leaving your IDE. Pick self-hosted if you need full control and have the infrastructure team to manage it.

Why Run Claude Code in the Cloud?

Running Claude Code locally ties up your machine and stops when you close the laptop. Running Claude Code in the cloud lets the agent work while you do other things, and running multiple agents in parallel multiplies throughput.

Common reasons to move Claude Code to the cloud:

• Run tasks in the background. Start a task and check on it later.
• Parallel execution. Run many Claude Code agents on different tasks simultaneously.
• Persistent environments. Keep development stacks running without tying up your local machine.
• Post-PR automation. Have agents respond to review comments, fix CI, and handle merge conflicts.
• Team workflows. Multiple team members can dispatch and monitor agents from one interface.

Option 1: Murmur

Murmur is Macroscope's platform for running many coding agents in parallel in the cloud. It is agent-agnostic: you can run Claude Code or Codex with your own API keys.

How Murmur runs Claude Code

1. Each agent gets a dedicated VM. Not a container. A full VM with your development stack running.
2. Bring your own Claude API keys. Murmur does not mark up Anthropic pricing. You pay Anthropic directly.
3. Orchestration built in. Start agents from a Director session, Slack, Linear, or scheduled Flights.
4. Post-PR automation. Agents respond to review comments, fix failing CI, rebase, and resolve merge conflicts.

Setup

1. Apply for access at /murmur-beta
2. Connect your GitHub organization
3. Configure your environment (dependencies, services, credentials)
4. Add your Anthropic API key
5. Start a Director session and spawn Claude Code agents

When to pick Murmur

• You want many Claude Code agents running in parallel
• You need post-PR automation (review comments, CI fixes, rebases)
• You want to run agents in your own GCP or AWS VPC
• You want to bring your own Anthropic keys without markup

Limitations

• In early preview. Apply at /murmur-beta.

Option 2: Cursor Background Agents

Cursor Background Agents run coding tasks in the cloud from within the Cursor IDE. They support Claude models and produce pull requests while you work on other things.

How Cursor Background Agents work

1. Assign a background task. From Cursor, describe what you want done.
2. Agent runs in the cloud. The task executes on Cursor's infrastructure.
3. PR opens when done. The agent produces a pull request you can review.
4. Resume from Cursor. Check progress or continue work from another Cursor window.

Setup

1. Install Cursor IDE
2. Subscribe to Cursor Pro or above
3. Configure environment.json for your project
4. Use the Background Agents feature to dispatch tasks

When to pick Cursor Background Agents

• You already use Cursor as your primary editor
• You want cloud tasks without leaving your IDE
• Fast switching between local and cloud contexts matters

Limitations

• Tied to Cursor. Not useful if your team uses other editors.
• Post-PR automation is more limited than Murmur.
• Environments are configurable but not full VMs.

Option 3: Self-Hosted Cloud VMs

You can spin up your own cloud VMs with Claude Code installed. This gives full control but requires infrastructure work.

How to set up self-hosted Claude Code

1. Create a VM. On GCP, AWS, Azure, or another provider.
2. Install dependencies. Node.js, Git, your development stack.
3. Install Claude Code. Via npm: npm install -g @anthropic-ai/claude-code
4. Configure credentials. Set ANTHROPIC_API_KEY in the environment.
5. Clone your repository. And set up the development environment.
6. Run Claude Code. claude starts the agent.

When to pick self-hosted

• You need full control over the environment
• Compliance requires VMs in specific regions or networks
• You have an infrastructure team to manage VMs
• You want to integrate with internal tooling

Limitations

• No orchestration. You manage agent lifecycle yourself.
• No post-PR automation. You handle review comments and CI manually.
• No parallel scaling without building it yourself.
• Operational overhead for VM management, updates, credentials.

Option 4: GitHub Codespaces

GitHub Codespaces can run Claude Code, but it is not designed for background agent workflows.

Limitations

• No orchestration or dispatch
• No parallel agent execution
• No post-PR automation
• Intended for interactive development, not autonomous agents

Codespaces can be useful for one-off Claude Code sessions, but for production agent workflows, Murmur or Cursor Background Agents are better fits.

Comparison: Claude Code Cloud Options

How does Claude Code Cloud relate to AI code review?

Running Claude Code in the cloud produces more PRs per developer. AI code review keeps pace with that output.

When agents open PRs faster than humans can review them, the review queue grows. Macroscope Code Review reviews every PR with full-repository context. Approvability auto-approves low-risk changes so humans focus on what needs judgment. Fix It For Me turns review findings into committed fixes.

The combination of cloud agents producing changes and AI code review validating them is what makes the software factory model practical.

Run Claude Code in the cloud with Murmur

Each agent gets its own VM with your development stack. Apply for early access.

Apply for access

Related Pages

• What Is an AI Software Factory?
• Best Devin Alternatives 2026
• What Is Cloud AI Coding Agent Orchestration?
• Best Cloud AI Coding Agents 2026
• What Are Background Coding Agents?
• Murmur FAQ

Frequently Asked Questions

How do I run Claude Code in the cloud?

You can run Claude Code in the cloud using Murmur (fleet orchestration with your own API keys), Cursor Background Agents (cloud tasks from Cursor IDE), or self-hosted cloud VMs with Claude Code installed. Murmur and Cursor provide orchestration; self-hosted requires building it yourself.

What is the best way to run Claude Code in the cloud?

For most teams, Murmur is the best option. It runs Claude Code agents in dedicated VMs with your own Anthropic API keys, provides orchestration for parallel agents, and includes post-PR automation. Cursor Background Agents are a good alternative for teams already using Cursor.

Can I use my own Anthropic API key with cloud Claude Code?

Yes. Murmur lets you bring your own Anthropic API key with no markup. Self-hosted VMs also use your own key directly. Cursor uses its own model access, though you may be able to configure custom keys depending on your plan.

How much does it cost to run Claude Code in the cloud?

You pay for compute (VM time) plus Anthropic API usage. On Murmur, you pay Anthropic directly at their standard rates plus Murmur's platform fee. On self-hosted, you pay your cloud provider plus Anthropic. On Cursor, costs are bundled into your Cursor subscription plus on-demand usage.

Can I run multiple Claude Code agents in parallel?

Yes, with Murmur or self-hosted infrastructure. Murmur is built for parallel agent execution. Self-hosted requires building the orchestration yourself. Cursor Background Agents also support parallel tasks.

What is Murmur?

Murmur is Macroscope's platform for running many coding agents in parallel in the cloud. It is agent-agnostic (Claude Code or Codex with your own keys), provides orchestration, and includes post-PR automation. Agent VMs can run in Murmur Cloud or in your own GCP or AWS VPC. Apply at /murmur-beta.

Is there a free way to run Claude Code in the cloud?

You can run Claude Code on your own cloud VMs and pay only for compute plus Anthropic API usage. There is no free tier for managed platforms like Murmur or Cursor Background Agents, but both have trial periods or credits for new users.

How do I set up Claude Code on a cloud VM?

Install Node.js and Git, run npm install -g @anthropic-ai/claude-code, set ANTHROPIC_API_KEY in your environment, clone your repository, and run claude. You will need to configure your development stack (databases, services, dependencies) the same way you would locally.