Skip to content

Pinggy-io/cli-js

Repository files navigation

Pinggy CLI

Create secure, shareable tunnels to your localhost and manage them from the command line.

Key features

  • HTTP, TCP, UDP, TLS, TLSTCP tunnels to localhost
  • SSH-style and user-friendly flags
  • Web debugger for HTTP tunnels
  • Extended options for auth, header manipulation, IP allowlists, CORS handling, etc.
  • Remote management via secure WebSocket connection (works with Pinggy Dashboard)
  • Configurable logging to file and/or stdout
  • Save and load configuration files
  • Simple file server mode for quickly sharing local files
  • Built-in TUI (Text User Interface) for viewing tunnel statistics, requests, and responses in real time

Requirements

  • Node.js 18+ (recommended). The CLI uses modern ESM and WebSocket features.
  • A network connection that allows outgoing WebSocket/HTTPS traffic.

Installation

Global install is recommended for system-wide "pinggy" command.

  • Using npm:
 npm install -g pinggy

After install, verify:

  pinggy --help

Quick start

  • Start a basic HTTP tunnel to localhost:3000:
 pinggy -R0:localhost:3000
  • Start a TCP tunnel (e.g., SSH on port 22):
  pinggy -R0:localhost:8000 tcp@free.pinggy.io
  • Start HTTP tunnel with web debugger on 4300:
pinggy -R0:localhost:8000 -L4300:localhost:4300

The CLI prints generated public URLs (HTTP/HTTPS or TCP) and keeps running until you press Ctrl+C.

Usage

Basic syntax: pinggy [options] [user@domain]

  • user@domain is optional. Domain can be any valid domain supported by the service backend (e.g., ap.example.com).

Options

The CLI supports both SSH-style flags and more descriptive long flags. Below is a consolidated list (only public ones are shown here). For the most up-to-date help, run pinggy --help.

Port Forwarding

Flag Description Example
-R, --R Local port forwarding (SSH-style) -R0:localhost:3000
-L, --L Web debugger address (SSH-style) -L4300:localhost:4300

Connection

Flag Description Example
-p, --server-port Pinggy server port (default: 443) --server-port 8080
--type Type of connection (e.g., tcp) --type tcp
-l, --localport Local endpoint [protocol:][host:]port --localport https://localhost:8000
-d, --debugger Port for web debugger -d 4300
--token Token for authentication --token abc123

Logging

Flag Description
--loglevel Logging level: ERROR, INFO, DEBUG
--logfile Path to log file
--v Print logs to stdout
--vv Detailed logs (Node.js SDK + Libpinggy)
--vvv Enable logs from CLI, SDK, and Libpinggy

Config

Flag Description
--saveconf <file> Create configuration file with provided options
--conf <file> Load configuration from file (CLI flags override)

File server

Flag Description
--serve <path> Serve files from a local directory via simple web server

AutoReconnect

Flag Description
--autoreconnect, -a Automatically reconnect tunnel on failure

Remote control

Flag Description
--remote-management <token> Enable remote tunnel management
--manage <addr> Remote management server (default: dashboard.pinggy.io)
--NoTUI Disable TUI in remote management mode

Misc

Flag Description
--version Print version and exit
-h, --help Show help and exit

Extended options

Extended options provide advanced controls. Specify them as positional values like x:https or w:192.168.1.0/24 alongside other CLI flags.

  • x:https Enforce HTTPS-only (HTTP redirected to HTTPS).
  • x:passpreflight | x:allowpreflight Allow CORS preflight to pass unchanged.
  • x:reverseproxy Disable built-in reverse-proxy header injection.
  • x:xff Add X-Forwarded-For.
  • x:fullurl | x:fullrequesturl Include original request URL.
  • w:[,...] Whitelist IPs (IPv4 CIDR).
  • k: Set Bearer token(s) for auth (repeatable).
  • b:user:pass Add Basic Auth credentials (repeatable).
  • a:Key:Val Add header.
  • u:Key:Val Update header.
  • r: Remove header.

Examples:

  • Enforce HTTPS and XFF for local HTTPS server on 8443: pinggy x:https x:xff -l https://localhost:8443

  • Allow only a local subnet: pinggy w:192.168.1.0/24 -l 8080

To generate advanced CLI arguments, use Configure from Pinggy.io

Remote management

You can control tunnels remotely using a secure WebSocket connection.

  • Start remote management with a token:
 pinggy --remote-management <API KEY>
  • Specify a management server (default is wss://dashboard.pinggy.io):
 pinggy --remote-management <API KEY> --manage wss://custom.example.com

Logging

You can control logs via CLI flags (which override environment variables). If logfile is provided, the log directory will be created if it does not exist.

  • To log to file and stdout at INFO level:
  pinggy -p 3000 --logfile ~/.pinggy/pinggy.log --loglevel INFO --v

If you provide --v, --vv, or --vvv without specifying a log level, the default log level is INFO.

Saving and loading configuration

  • Save current options to a file:
  pinggy -p 443 -L4300:localhost:4300 -t -R0:127.0.0.1:8000 qr+force@free.pinggy.io   x:noreverseproxy x:passpreflight x:xff --saveconf myconfig.json
  • Use a config as base and override with flags:
pinggy --conf ./myconfig.json -p 8080

File server mode

Serve a local directory quickly over a tunnel: pinggy --serve /path/to/files Optionally combine with other flags (auth, IP whitelist) as needed.

Signals and shutdown

Press Ctrl+C to stop. The CLI traps SIGINT and gracefully stops active tunnels before exiting.

Versioning

This package follows semantic versioning. See package.json for the current version.

License

Apache License Version 2.0

About

Pinggy CLI for Node.js

Resources

License

Stars

Watchers

Forks

Packages

No packages published

Contributors 2

  •  
  •