Install
On macOS or Linux:
Terminal
curl -fsSL https://techbeaver.io/install.sh | shOn Windows, in PowerShell:
PowerShell
irm https://techbeaver.io/install.ps1 | iexThe installer downloads the release for your system, checks it against the published checksum, and refuses to install anything that does not match.
Sign in
Run beaver auth login. Your browser opens on a Techbeaver page listing what the CLI is asking to do; approve it and the terminal says you are signed in.
On a server or any machine without a browser, run beaver auth login --no-browser. It prints a short code and a link: open the link on your phone or laptop, enter the code, and the terminal finishes signing in on its own.
Tell it which app you mean
Put a beaver.toml file in your project folder naming the project and the app:
beaver.toml
project = "storefront"app = "storefront"Every command run in that folder now acts on that app, so you never retype it.
Deploy and watch it go live
beaver apps deploy --wait starts a deploy and follows the build log until the app is live, then prints its URL. Pressing Ctrl-C stops watching, not the build. If your repository has a Dockerfile at its root, it is used; otherwise the build plan is generated from your code.
Logs and settings
beaver apps logs --followstreams your app’s logs as they happen.beaver apps env set LOG_LEVEL=infochanges one environment variable and leaves the others untouched. A running app rolls out with the new value on its own.
Create a managed PostgreSQL database
beaver plans list --db lists the database plans with their IDs. Then:
Terminal
beaver db create orders-db --plan <plan-id> --billing-cycle monthlyIt answers with a payment link for the plan. The database is created once you pay, and the CLI never pays on its own.
Deleting needs a person
beaver apps delete does not delete anything by itself, even with --yes. It prints an approval link and says what will go; the delete happens only after someone signed in approves it in a browser.
Why
Use it from an AI client on your machine
beaver mcp runs the CLI as a local MCP server, so an AI client on your machine can use the same tools with your own credential. To connect a client that runs in the cloud instead, add https://mcp.techbeaver.io/mcp as a connector.
