Skip to main content
The Hyperline CLI lets you manage your billing operations directly from the terminal. It’s designed to be agent-first — optimized for AI agents like Claude Code, Cursor, Windsurf, and any tool that can run shell commands.

​
Why a CLI for AI agents?

AI agents work with tools to accomplish tasks. Today, most agents use MCP (Model Context Protocol) to connect to external services. While MCP works well, it has limitations when used at scale:
  • Context bloat — MCP servers expose dozens of tools, each with full schemas that consume valuable context window space.
  • Slower on complex tasks — For multi-step workflows, agents using MCP need to discover tools, read schemas, and format structured requests for each operation. On complex questions, MCP agents can loop for minutes trying to find the right tool combination.
CLIs solve these problems:
  • Token-efficient — Short commands like hyperline customers list are natural for a text-predicting model to write and require minimal context to understand.
  • No tool discovery overhead — Agents don’t need to load a catalog of tools or read JSON schemas. They just run commands and read the output.
  • Composable — CLI output can be piped, filtered with jq, or chained with other commands — patterns that agents handle natively.

​
Installation

Verify the installation:

​
Authentication

​
Browser login

The simplest way to authenticate — opens your browser to securely log in with your Hyperline account:
Credentials are stored locally in ~/.hyperline/. To log out:

​
API key

For server environments, CI/CD pipelines, or AI agents running without a browser:
API keys can be generated from your account Settings.
Keep your API key secret. Do not commit it to version control — use environment variables or a secrets manager instead.
For sandbox environments, set the base URL via environment variable or flag:

​
Company selection

If you have access to multiple companies, the CLI uses the last used company by default. To switch to a specific company:

​
Usage

The CLI follows a consistent hyperline <resource> <action> pattern:

​
Available resources

Use --help on any command to see available actions and options:

​
Output formats

By default, the CLI outputs human-readable text. Use --output json for structured output:

​
Global options

​
Setting up for AI agents

​
Claude Code

Add the Hyperline CLI to your project’s CLAUDE.md so Claude knows it’s available:

​
Cursor / Windsurf

Add instructions to your project rules (.cursor/rules/ or .windsurfrules) so the agent knows to use the CLI:

​
Any agent with shell access

Any AI agent that can execute shell commands can use the Hyperline CLI. Set the HYPERLINE_API_KEY environment variable in the agent’s environment, and the CLI is ready to use — no additional configuration or tool registration needed.

​
CLI vs MCP — which should I use?

You can use both. The MCP server is great for conversational AI assistants, while the CLI shines for agents that work in the terminal.

Hyperline MCP Server

Learn how to set up the Hyperline MCP server for AI assistants without terminal access.