LinkSocks.js

A SOCKS proxy over WebSocket service built on Cloudflare Workers, designed for simple intranet penetration with no server setup required.

If you own certain intranet server, please use FRP or Cloudflare Tunnel.

This project is mainly used to quickly connect to uncertain user intranet environment.

中文文档 / Chinese README

Installation

Client Setup

1. Download the linksocks client from https://github.com/linksocks/linksocks/releases.
2. Select the appropriate version for your OS (Windows, Linux, macOS).
3. Extract and add to your PATH or run directly.

Server Setup (Optional)

This repository contains the Cloudflare Worker server-side code. You can:

1. Use our public server: https://l.zetx.tech

2. Deploy your own:

   

Usage

# Step 1: On machine A (inside the network you want to access)
linksocks provider -u l.zetx.tech -c your_token

# Step 2: On machine B (where you want to access the network)
linksocks connector -t your_token -u l.zetx.tech -p 1180

After running both commands, you can access the internal network through the SOCKS5 proxy at 127.0.0.1:1180 on machine B.

Configure your browser or applications to use this SOCKS5 proxy to access internal network resources.

Key Benefits

• Zero Infrastructure: No need to maintain servers, infinite bandwidth.
• Plug and Play: Just two commands to establish connection, no config file, no online setup.
• Proxy at Cloudflare Edge: Low latency for global application.
• Load Balancing: Multiple provider using same token will share the load.

Note

This service uses Cloudflare Durable Objects which has free tier limitations (13,000 GB-s / day). For more information on limits and pricing, see: https://developers.cloudflare.com/durable-objects/platform/pricing/

Health Check API

The worker exposes a health status API for external monitoring. Health results are cached for 1 hour in the Durable Object and refreshed when the cache is close to expiring.

GET /api/health

Returns the current cached health status. If no cache exists or it has expired, a fresh probe is run automatically.

Status codes:

• 200 — healthy (all checks passed)
• 503 — unhealthy (one or more checks failed)

Response fields:

Field	Type	Description
active	boolean	true when all checks pass
checkedAt	number	Timestamp (ms) when the check ran
cacheExpiresAt	number	Timestamp (ms) when the cache expires
refreshAfter	number	Timestamp (ms) after which a poll request will trigger a refresh
origin	string	Worker origin used for the self-request probe
checks	object	Individual check results (see below)
usage	object|null	Current-day Durable Object usage from GraphQL, or null if unavailable
reason	string|null	Semicolon-separated failure reasons, or null if healthy
cached	boolean	Whether this response was served from cache
refreshed	boolean	Whether a fresh probe was run for this response
cacheAgeMs	number	Milliseconds since the cached check ran (0 if freshly refreshed)

checks object:

Field	Type	Description
durableObjectAccess	boolean	Successfully called a method on the Token Durable Object
storageRoundTrip	boolean	Wrote a value to DO storage and read it back
workerRequest	boolean	Worker fetched its own /api/health/ping endpoint
storedDataWithinLimit	boolean	storedBytes is below 5 GB

Example:

curl https://your-worker.example.com/api/health

GET /api/health/poll

Polling endpoint designed for external uptime monitors. Behaves the same as /api/health, but additionally checks whether the cache is close to expiring (within 10 minutes). If so, a fresh probe is run before returning.

Typical usage: configure your monitoring tool to poll every 5 minutes. The cache will auto-refresh once per hour, and near-expiry polls will keep it fresh.

curl https://your-worker.example.com/api/health/poll

Force Refresh

Append ?force=1 to skip the cache entirely and run a fresh probe immediately:

curl https://your-worker.example.com/api/health/poll?force=1