Keep your Claude Code working when your laptop is shut

Let it take tasks from your phone, on your own Claude login.

You've become its babysitter

You leave the lid open, check on it before bed and hope it makes it through the night.

Give it a computer of its own

Same Claude Code, same login, on a computer that keeps going when you close the lid. Your laptop goes back to being yours, and your agent works the hours the work needs.

Hand it work in the evening, review it over coffee

Before you log off, hand it the job you were saving for a free day. It works through the night on its own computer and leaves a branch and a short note by the time you open your laptop.

When the work is done it sleeps, so you don't pay for a server idling all night. Your next message wakes it with every file where it left them.

Work you couldn't give it before

Once it doesn't need you to keep it awake, people stop handing it one task at a time. They give it routines: upgrade dependencies every night, read new bug reports before standup, fix main when CI breaks.

Each routine runs on its own and posts what it did in the channel you pick, so the day starts with finished work to review.

Send it work from anywhere

You can message it from the Laze iPhone app or any browser, the way you'd message a colleague: a failing test handed over from the train, or a question about what changed on main before a meeting.

It answers in the same chat, so the thread is the record of what it did.

It runs on the Claude plan you already pay for

Sign in once with your Claude account, and the work counts against your Pro or Max plan. No API key, and no second bill for tokens.

It's the same Claude Code with your settings, skills and CLAUDE.md, and only your own agents can use your login.

It picks up where it left off

Its computer keeps your checkouts, packages and build caches between tasks, so the next test run starts warm instead of from a fresh clone.

Connect GitHub and it works on your repos and pushes branches for you to review.

Who runs it and where your code lives

Laze is made by Reasoning Engine B.V., a company in the Netherlands, and runs in the European Union. Your agents work on your own computer, a separate machine with no path to Laze's database.

What you write in Laze is stored encrypted with a key for your workspace, and Laze never trains AI models on it. The security page explains the details, and the docs cover every step of the setup.

Mara's computer
  • Runs in the European Union
  • Its own machine, no path to Laze's database
  • Your Claude login, used by your agents only
  • Never used to train AI models

Set it up in a few minutes

There's no server to rent and no SSH keys to copy around.

1

Make an agent

Create a free Laze account and pick Claude Code for your agent.

2

Sign in with Claude

Sign in once with your Claude account from its computer.

3

Connect GitHub

Give it the repos it should work on.

4

Send the first task

Message it from your phone or a browser, and close the laptop.

Pricing

Every plan, Free included, comes with a Small computer: 1 vCPU, 2 GB of memory and 20 GB of disk. Your Claude usage comes out of your Claude subscription, and a bigger computer for heavier builds costs 10 to 70 USD a month more. Laze is free for you and one colleague, and the Team plan starts at 20 USD a month when a third person joins.

Questions people ask

Is it the real Claude Code?

Yes. It's the same Claude Code CLI you run in your terminal, signed in with your own Claude account.

Does it run around the clock?

It runs whenever there is work, and a task in progress keeps it awake until it's done. Once idle, the computer sleeps until a message or a schedule you set wakes it.

Who else can use my Claude login?

No one. Your login works only for your own agents, and your computer runs no one else's.

Can I run tasks on a schedule?

Yes. You can ask it to check something every morning or run a job every Friday, and the schedule wakes the computer for it.

How does it get to my code?

Connect GitHub, and it works on your repos and pushes branches for you to review.

Can my team message it too?

Yes. Add a teammate or a channel to the agent, and they can message it directly, without a Claude account of their own.

Close the laptop and let it work

Start for free