Built IPv6 Proxy Multiplexer - One IPv4 VPS → Thousands of Unique IPv6 Endpoints

Hey BHW developers,​

• Install on Linux VPS with IPv6 /48 or /64 subnet
• Creates HTTP proxies on different ports via REST API
• Each proxy port automatically gets assigned a unique IPv6 address
• To remote services, every request appears to come from a completely different client across the internet
• Self-hosted, so you own the entire infrastructure

# Create 100 proxies instantly
POST /api/proxies/8000-8099

# Each proxy gets unique IPv6 from your subnet
# 8000 → 2a13:d207:8f::a1b2
# 8001 → 2a13:d207:8f::c3d4
# etc.

# Export standard format
GET /api/export/proxies.txt
# Output: IP:PORT:USER:PASS

• VPS providers give you huge IPv6 blocks (up to 281 trillion addresses)
• Most people just let them sit unused
• Proxy services charge monthly for what you could self-host
• Each request from different IPv6 = massive unique IP pool

• Go binary with SQLite persistence
• IPv6 binding to network interface
• REST API for management
• Recovers state after restart
• Authentication per proxy

1. Is this approach technically sound? Any network/IPv6 experts see issues?
2. Real-world applications? Where would unlimited unique IPs be valuable?
3. Performance concerns? Running hundreds/thousands of proxy processes on one server?
4. Feature gaps? What would make this more useful for automation/scraping?

tapestry6637 said:

Hey BHW developers,​

Been working on an interesting technical challenge and wanted to get feedback from the community. Built a system that turns a single IPv4 VPS into thousands of unique proxy endpoints using IPv6 subnet allocation.

The technical concept:

• Install on Linux VPS with IPv6 /48 or /64 subnet
• Creates HTTP proxies on different ports via REST API
• Each proxy port automatically gets assigned a unique IPv6 address
• To remote services, every request appears to come from a completely different client across the internet
• Self-hosted, so you own the entire infrastructure

Example workflow:

Code:

# Create 100 proxies instantly
POST /api/proxies/8000-8099

# Each proxy gets unique IPv6 from your subnet
# 8000 → 2a13:d207:8f::a1b2
# 8001 → 2a13:d207:8f::c3d4
# etc.

# Export standard format
GET /api/export/proxies.txt
# Output: IP:PORT:USER:PASS

Why this approach:

• VPS providers give you huge IPv6 blocks (up to 281 trillion addresses)
• Most people just let them sit unused
• Proxy services charge monthly for what you could self-host
• Each request from different IPv6 = massive unique IP pool

Technical implementation:

• Go binary with SQLite persistence
• IPv6 binding to network interface
• REST API for management
• Recovers state after restart
• Authentication per proxy

Questions for the community:

1. Is this approach technically sound? Any network/IPv6 experts see issues?
2. Real-world applications? Where would unlimited unique IPs be valuable?
3. Performance concerns? Running hundreds/thousands of proxy processes on one server?
4. Feature gaps? What would make this more useful for automation/scraping?

Currently it's just REST API - considering adding web dashboard, Docker containers, monitoring, etc.

For context: I'm not trying to sell anything here, genuinely curious about the technical merit and whether this solves real problems people have. The IPv6→proxy mapping concept seems underutilized.

Appreciate any technical feedback, architecture suggestions, or thoughts on practical applications!

---

P.S. - If you're curious about the implementation details or want to discuss the IPv6 networking aspects, happy to share more technical specifics.

Click to expand...

Liddells Doll said:

This is basically NAT.
The main issue with your idea is that HTTP is a little slow and possibly overkill for your use-case.
Oh, and if you bind a port for each proxy endpoint you will have a hard limit of 65535 concurrent connections.

Click to expand...