No: your agent gets authorized in your browser, once. For a CI without a browser, a token takes its place: your agent creates it, and you confirm it in one click.

Does my agent see my keys?

No: your keys stay in while, and calls to an external API go through while, which adds the key.

What can my agent do without me?

It can deploy, and read the logs and the database. Sensitive actions, like deleting a production database, wait for your approval.

Connect while to Claude Code over MCP, in one command

Claude Code already writes your app. Connected to while over MCP, with one command, it also deploys it, with its database and who can open it, and asks for your approval on what matters.

Prerequisites

Claude Code, installed on your computer.
A browser, to authorize your agent once.

Step 1: Connect while to Claude Code

In a terminal, run this command¹:

Terminal

claude mcp add --transport http --scope user while https://app.while.sh/mcp

Or let your agent do it: paste “Connect to while with my account (while.sh/agents), then deploy my app.” into it, and it connects itself.

Step 2: Authorize it in your browser

On the first call to while, a page opens in your browser. If it doesn’t open, type /mcp in Claude Code, choose while, then Authenticate. On that page, choose “A new app”, then click Authorize.

What your agent does

Claude Code then sees while’s tools: deploy, create a database, read the logs, query the database, open a preview, roll back. It asks for your approval before calling a tool, and you can allow the read-only ones up front.

Deploy your app

You, to your agent: Deploy my app with while.

Your agent writes while.yaml, sends the code, and gives you your app’s address, over HTTPS. The details are in the guides, from Next.js to Django.

Understand an error

You, to your agent: Why has the app been returning errors since this morning?

Your agent reads the logs and the grouped errors in while, finds the cause and fixes it.

Query your database

You, to your agent: How many people signed up this week?

Your agent reads your database in SQL over MCP, with your permissions on the data, and it can’t write to it.

See a new version before you deploy it

You, to your agent: Show me the new version before you deploy it.

Your agent opens a preview, at its own address, with its own data. You look at it, then your agent puts it into production.

Go back to the previous version

You, to your agent: Go back to the previous version.

In the 30 minutes after a deployment, the previous version comes back in a second. After that, it takes a few seconds. Once the previous version is back, your agent reads what broke in while’s logs, fixes it and deploys again.

Connect all the agents on your computer

You, to your agent: Connect while to all my agents.

One command connects Claude Code, Codex, Cursor, Gemini CLI and VS Code at once. Read the script first, if you like:

Terminal

curl -fsSL https://app.while.sh/install.sh | sh

In the docs: Previews and Versions and rollback.

Connect while to Claude Code over MCP to deploy your apps