Skip to main content
Every command is generated from the same OpenAPI document that serves the API, so a command exists here only if the endpoint behind it exists. Commands are named after their path: GET /v1/stats/overview is traceten stats overview. Each command takes the flags listed below plus the global flags. A flag is the endpoint’s parameter with underscores replaced by dashes, so site_id is --site-id (aliased to --site). What each parameter means, and what comes back, is on the matching API reference page. A flag marked (JSON) takes a JSON literal, quoted for your shell: --steps '[{"kind":"pageview","match":"exact","value":"/pricing"}]'. A flag marked (flag) is a boolean and takes no value. Everything else is a plain string or number. Which permission a command needs is the endpoint’s permission. The maintained table is Which endpoints need which permission, and a 403 names the one you are missing in required_scope. Commands marked --yes refuse to run without it. They are the same operations the MCP server gates behind a confirmation, plus the ones the MCP does not expose at all.

account

ai-crawls

annotations

breakdowns

detection

events

funnels

goals

integrations

revenue

sessions

sites

sources

stats

team

visibility

visitors

webhooks

Optional flags, by command

Next