Resident proxies via your own VPS: parsing without bans

Creating residential proxies yourself is possible by renting several budget VPS in various geographic locations and configuring specialized software (3proxy, Squid, or gost) for IP rotation. This allows you to reduce scraping costs from $500/month for commercial packages to $50/month for your own infrastructure of 10-15 nodes.

Why are DIY residential proxies more profitable than ready-made solutions?

Commercial providers, such as Bright Data, Oxylabs, or Smartproxy, resell home user traffic, charging per gigabyte or per port. With large scraping volumes, bills can reach thousands of dollars. Building your own network based on virtual servers allows you to pay only for power rental while getting unlimited traffic.

Scraping Economics: Cost Comparison

A large project for data collection from marketplaces or social networks requires a high frequency of requests from different IPs. If you use scraping proxies from major providers, the price per 1 GB of traffic varies from $3 to $15. At the same time, renting a VPS with a 1 Gbps channel and traffic from 1 to 20 TB costs a fixed amount.

Technical Advantage of Infrastructure Ownership

When you set up DIY residential proxies, you gain access to low-level TCP stack and HTTP header settings. This is critical for bypassing protection systems like Cloudflare or Akamai, which analyze not only IPs but also TLS fingerprints. On your own server, you can spoof MTU, change TTL, and configure specific ciphers so that traffic looks like a regular user request.

How to organize a proxy chain VPS for maximum anonymity?

Using a single VPS will quickly lead to its IP address being blocked by the target resource. To implement anti-ban scraping, you need to build a proxy chain where the entry node distributes requests among dozens of exit nodes in different data centers.

Star Architecture with a Central Balancer

In this scheme, you have one "Entry Node" to which your script sends all requests. This server, using rotating proxy logic, redirects the request to one of the "Exit Nodes". Thus, the target site sees IP addresses of different servers scattered around the world.

• Entry node: accepts HTTP/SOCKS5 traffic.
• Exit nodes: 10-50 cheap VPS in different countries.
• Rotation: random selection of an exit node for each new request or session.

To minimize infrastructure costs, it's worth considering an AWS EC2 alternative for backend, where you can rent servers with pay-as-you-go resources without overpaying for the brand.

Choosing Locations and Providers for Effective Scraping

The key to success when creating your own proxy network is subnet diversity. If all 10 of your servers are in the same data center (e.g., Hetzner or DigitalOcean), the anti-fraud system will ban the entire subnet. You need to choose providers that offer IPs from different Autonomous Systems (AS).

Recommended VPS Specifications for a Proxy Node:

1. CPU: 1 vCPU (proxy servers consume few CPU resources).
2. RAM: 512 MB - 1 GB RAM (sufficient for 3proxy or Squid).
3. Disk: 10 GB SSD (it's better to store logs in RAM or delete them).
4. Network: 1 Gbps Shared.
5. OS: Debian 11/12 or Ubuntu 22.04.

When searching for exit node locations, it's useful to explore DigitalOcean alternatives, as small local providers often have "clean" IP addresses that haven't yet landed on the spam lists of major platforms.

Step-by-Step 3proxy Setup for Creating Your Own Network

3proxy is a cross-platform open-source proxy server that is ideal for creating DIY residential proxies due to its lightweight nature and flexible configuration.

Installing 3proxy on Debian/Ubuntu

apt update && apt install build-essential -y
wget https://github.com/z3APA3A/3proxy/archive/refs/tags/0.9.4.tar.gz
tar xzf 0.9.4.tar.gz
cd 3proxy-0.9.4
make -f Makefile.Linux
mkdir /etc/3proxy
cp bin/3proxy /usr/bin/

Exit Node Configuration

On each exit server, create the file /etc/3proxy/3proxy.cfg. We will set up IP-based authorization (your entry node's IP) so that outsiders cannot use your proxy.

daemon
nserver 8.8.8.8
nserver 1.1.1.1
nscache 65536
timeouts 1 5 30 60 180 1800 15 60
users user1:CL:password123
# Allow access only from the entry node IP
allow * 1.2.3.4 * * 
auth strong
proxy -p3128 -n -a
socks -p1080 -n -a
flush

Here, 1.2.3.4 is the IP of your central server. Now this node is ready to accept traffic and forward it.

Anti-ban Scraping Setup: Rotation Logic on the Entry Node

The hardest part is making the entry node constantly change exit IPs. For this, you can use Squid or the same 3proxy with the parent function. A rotating proxy allows your scraper to connect to a single port and receive a new address every time on the output.

Example of Rotation Configuration in 3proxy

# List of exit nodes
parent 1000 socks5 11.22.33.44 1080 user1 password123
parent 1000 socks5 55.66.77.88 1080 user1 password123
parent 1000 socks5 99.00.11.22 1080 user1 password123

# Selection logic: random
proxy -p8080 -a

The parent parameter with equal weight (1000) forces 3proxy to distribute requests among the specified servers. If one of the servers goes down, 3proxy will temporarily exclude it from the chain, ensuring high system fault tolerance.

Using GOST to Build Advanced Tunnels

GOST (GO Simple Tunnel) is a powerful tool written in Go that allows creating complex proxy chain VPS using various protocols: HTTP2, QUIC, KCP. This is especially useful if standard proxy protocols are blocked by the provider or the target resource.

Advantages of GOST for Scraping:

• Support for multi-hop chains (A -> B -> C).
• Ability to encapsulate traffic in TLS to bypass DPI.
• Built-in load balancing with node health checks (healthcheck).

To manage such complex configurations, it's convenient to use graphical interfaces. For example, the 3x-ui panel allows you to quickly set up Reality and VLESS protocols, which are virtually indistinguishable from regular HTTPS traffic.

# Running GOST in rotator mode on the entry node
gost -L=:8080 -F=socks5://user:pass@exit1:1080?weight=1 -F=socks5://user:pass@exit2:1080?weight=1

Optimizing the Network Stack for High-Load Scraping

When you run DIY residential proxies, standard Linux settings can become a bottleneck. When scraping in 500-1000 threads, the server will quickly exhaust the limit of open files and ports.

Configuring sysctl.conf

Add the following parameters to /etc/sysctl.conf to optimize network connection handling:

# Increase open files limit
fs.file-max = 1000000

# Speed up TCP port reuse
net.ipv4.tcp_tw_reuse = 1
net.ipv4.ip_local_port_range = 1024 65535

# Increase connection queue
net.core.somaxconn = 65535
net.ipv4.tcp_max_syn_backlog = 65535

After making the changes, run sysctl -p. This will allow a single VPS to handle tens of thousands of simultaneous proxy requests without delays.

Use Cases: From SEO to Sneaker Bots

Your own VPS-based proxy network is versatile. It is suitable for:

1. Collecting search engine rankings (Google, Yandex).
2. Monitoring competitor prices on marketplaces (Amazon, Wildberries).
3. Automating actions on social networks.
4. Working with specialized software. For example, setting up a Sneaker bot on VPS requires minimal latency (ping), which is easier to achieve on your own nodes in the required location.

Security and Protection of Your Proxies

Open proxy servers are quickly discovered by scanning bots and end up on public lists. To ensure your DIY residential proxies don't become public property, follow these security rules:

• Always use strong authentication (login/password).
• Limit access to proxy ports via iptables or ufw, allowing connections only from your working IPs.
• Change standard ports (3128, 8080, 1080) to random ones (e.g., 47291).
• Regularly update software (3proxy, Squid) to patch vulnerabilities.

Example of UFW Firewall Setup

ufw default deny incoming
ufw allow ssh
# Allow entry node
ufw allow from 1.2.3.4 to any port 47291
ufw enable

Conclusions

Creating your own residential proxy network based on VPS is the most effective way to scale scraping while significantly reducing costs. For best results, use a combination of 10-15 servers in different data centers and set up rotation via 3proxy or GOST.