Skip to main content
The CLI wraps all 62 Supersonic tools as shell commands. It outputs pretty tables for humans and --json for AI agents.

​
Install

npx supersonic-cli login
Or install globally: 
npm i -g supersonic-cli
supersonic login

​
Authenticate

supersonic login gives you two options:
  1. Login with browser — opens supersonic.cv, creates an API key automatically
  2. Paste an API key — from your dashboard
You can also pass the key directly: 
supersonic login --key supersonic_live_YOUR_KEY

​
Commands

# Schema
supersonic objects list
supersonic objects get contacts
supersonic objects create Deals
supersonic objects add-field deals value currency

# Records
supersonic records list contacts
supersonic records create contacts --set name=Jane,email=jane@acme.com
supersonic records update <id> --set stage=qualified
supersonic records search "Acme"
supersonic records bulk-create contacts -f data.json

# Pipelines
supersonic lists list
supersonic views data <view-id>          # auto-renders kanban for board views

# Tasks
supersonic tasks create "Follow up with Jane" --due 2026-04-01
supersonic tasks list --mine

# Analytics
supersonic analytics records deals --metric count --group-by stage

​
Output

Every command supports three output modes: 
supersonic records list contacts          # table
supersonic records list contacts --json   # JSON
supersonic records list contacts --csv    # CSV
Data goes to stdout, status messages go to stderr. This means you can pipe output: 
supersonic records list deals --json | jq '.items[].data.name'

​
Environment variables

VariableDescriptionDefault
SUPERSONIC_API_KEYAPI key (overrides config file)—
SUPERSONIC_ENDPOINTAPI endpointhttps://mcp.supersonic.cv

​
How AI agents use the CLI

An AI coding agent (Claude Code, Codex) runs CLI commands as bash calls instead of loading MCP tool schemas into its context window. This uses 10–30x fewer tokens per operation. 
Agent thinks: "I need to find the user's contacts"
Agent runs:   supersonic records list contacts --json
Agent reads:  [{"id": "...", "data": {"name": "Jane", ...}}, ...]