Skip to main content

Documentation Index

Fetch the complete documentation index at: https://docs.bentolabs.ai/llms.txt

Use this file to discover all available pages before exploring further.

bentolabs-cli is a command-line client for Bento. Use it to list traces, inspect signals, run analytics, and hit any API endpoint from your terminal.

​
Requirements

Python 3.10 or higher.

​
Install

With uv, which installs the CLI into an isolated tool environment: 
uv tool install bentolabs-cli
Or with pip: 
python3 -m pip install --user bentolabs-cli
Confirm the install: 
bentolabs version

​
Sign in

Authenticate through your browser: 
bentolabs auth login
Your browser opens a Bento Allow / Deny page. Click Allow and the CLI is signed in as you. Tokens are saved in the OS keychain (macOS Keychain, Linux Secret Service, Windows Credential Locker). On headless Linux without secret-service, the CLI falls back to a chmod 0600 file. Tokens refresh automatically before they expire. 
bentolabs auth whoami   # see who is signed in
bentolabs auth logout   # revoke the session and clear tokens

​
Pick a default workspace

CLI commands act on one workspace at a time. Set a default once so you don’t pass --workspace every time: 
bentolabs workspaces list           # see workspaces you belong to
bentolabs workspaces use <id>       # save this as the default
Override per command with the flag: 
bentolabs traces list --workspace <other-id>
The flag wins over the saved default.

​
Run your first command

List the five most recent traces in your default workspace: 
bentolabs traces list --limit 5 --output table
You should see a table of recent traces in your default workspace. If you see a hint about no workspace selected, run bentolabs workspaces use <id> first.

​
Next

Commands

Built-in commands, output modes, and how arguments map.

Python SDK

Send traces and signals from your application code.