warp

This library provides convenient access to the warp REST API from the command line.

The full API of this library can be found in api.md.

Contents

• Installation
• Usage
• API Reference
• Shell Completion
• Manual Pages
• Authentication
• Errors
• Client Options
• Retries and Timeouts
• Helpers
• Logging
• Requirements

Installation

# Homebrew — standalone binary, no Node.js required
brew install TeamWarp/tap/warp

Usage

warp [resource] [command] [flags]

warp benefits:health-plans list --api-key "$WARP_API_KEY" --limit 'limit' --statuses '["active"]'

The examples in the following sections assume a client configured as shown above.

See the API reference for every available operation.

Shell Completion

warp completion <shell> prints a completion script for bash, zsh, and fish. Add the matching line to your shell startup file to complete commands, subcommands, and flags with Tab.

# bash (~/.bashrc)
eval "$(warp completion bash)"

# zsh (~/.zshrc)
eval "$(warp completion zsh)"

# fish (~/.config/fish/config.fish)
warp completion fish | source

Manual Pages

Installing the package globally also installs man pages. man warp lists every command, and each command has its own page named after the command with spaces and : replaced by -.

man warp
man warp-<resource>-<command>

Authentication

Pass credentials to the generated client constructor. Environment variables are read automatically when supported by the target runtime.

Option	Type	Default	Description
--api-key	string | provider	-	Credential for the apiKey scheme. Defaults to WARP_API_KEY.

Declared schemes:

• apiKey API key in header x-api-key

Errors

Failed requests print a structured error to standard error and exit with a status that identifies the failure class. The error body carries the API's own message plus a stable code, the HTTP status, the requestId, and — where one applies — an actionable hint. Usage errors (exit 2) are reported as a plain message instead, since no request was made. Exit statuses: 0 success, 1 error, 2 usage, 10 auth-failed, 11 not-found, 12 rate-limited, 13 client-error, 14 server-error, 15 connection-error.

Documented error statuses: 400, 401, 403, 404, 409, 422, 429, 500.

Client Options

Configure the generated client by setting any of these options when you create it.

Option	Type	Default	Description
--base-url	<url>	-	Override the base URL for API requests.
--timeout	<ms>	-	Request timeout in milliseconds.
--max-retries	<count>	-	Number of retries for retryable failures.
--debug	flag	-	Enable SDK debug logging.

Retries and Timeouts

Generated clients support request timeouts and retry temporary failures such as network errors, 408, 409, 429, and 5xx responses. Retry delays honor Retry-After headers when present. Tune the retry and timeout client options shown above, or override them per request.

Helpers

• --format <format> — output format: auto, json, jsonl, pretty, raw, toon, or yaml.
• --format-error <format> — error output format: auto, json, jsonl, pretty, raw, toon, or yaml.
• --format toon — token-efficient structured output for AI agents; uniform lists collapse into one header plus a row per item, with a definitive item count.
• --transform <path> and --transform-error <path> — dot-path transform for data/error output.
• --raw-output, -r — print transformed string values without JSON quotes.
• --max-items <count> — bound iterator, streaming, and WebSocket command output.
• Errors carry a stable code and an actionable hint beside the API's own message, and each failure class exits with its own status: 1 error, 2 usage, 10 auth-failed, 11 not-found, 12 rate-limited, 13 client-error, 14 server-error, 15 connection-error.

Logging

• Pass --debug to any command to enable SDK debug logging on stderr.

Requirements

• None — the standalone binaries bundle their own runtime.

Powered by Scalar.