Crawler mixed pool
About 95% residential and 5% datacenter IPs for faster general scraping.
Global rotating proxy
Two proxy pools for speed and residential identity
Connect through one fixed gateway and choose between a faster crawler pool or a pure residential pool. Exits are globally random by default; broad region presets are optional when extracting proxies.
Crawler pool: about 95% residential + 5% DC100% residential poolSESSION on the pure residential poolTraffic or concurrency plans
Global rotating proxyready
TWO GLOBAL POOLS123PROXY NETWORK
PRODUCT PROFILECrawler pool + pure residential poolOne gateway, two routing profiles
READY
CRAWLER~95% RES + ~5% DC
PURE RES100% residential
LOCATIONGlobal random
SESSIONPure pool only
Client123ProxyTarget
2 poolsCrawler / pure residential
~5% DCFaster crawler mix
SESSIONPure pool support
Traffic / threadsTwo capacity models
01 / Product capability
Choose the proxy pool first, then size the request model
The crawler pool usually provides faster public web collection by mixing a small share of datacenter IPs into a residential-majority pool. The pure residential pool keeps residential network identity and supports SESSION.
Pure residential pool
A pool containing only residential IP exits.
SESSION support
Use SESSION on the pure residential pool to keep an exit for multi-step requests.
Global random by default
Use global random exits directly, or choose Europe and America, North America, Europe, Asia, United States, or Japan and Korea when extracting proxies.
Traffic plans
Pay for transferred data when workloads are elastic or project-accounted.
Concurrency plans
Pay for in-flight request capacity with unmetered traffic in the plan.
02 / Plans & billing
Pay by transferred traffic or concurrent requests
Traffic plans fit variable workloads and measured data budgets. Concurrency plans fit sustained code-driven requests and do not meter transferred GB within the selected thread capacity.
Traffic or concurrency
Test real response sizes and latency before choosing. Browser pages can consume many concurrent requests for HTML, images, CSS, JavaScript, fonts, and APIs.
2 MODELSTraffic GB or concurrent threads
View proxy pricing
- Pool choices
- Crawler / pure residential
- Exit region
- Global random by default
- Traffic option
- Metered by GB
- Thread option
- Unmetered traffic
Planning note
One browser page is not one concurrent request
A browser can open 10–20 or more in-flight requests while loading a page. Concurrency packages are usually easier to model for requests, Scrapy, and direct HTTP clients.
03 / Workloads
Best for global random code-driven scraping
Choose this product for globally random exits or broad regional routing through a standard rotating gateway.
Public page crawling
High-frequency list, detail, and search result requests across public sites.
Residential multi-step requests
Pagination and short workflows that need a residential exit and SESSION.
Sustained request queues
Long-running code workers using a predictable concurrent request allowance.
04 / How it works
Connect once and let the gateway rotate exits
Select a pool, configure authentication, choose a billing model, and send requests through the fixed gateway.
01
Select a proxy pool
Use the crawler pool for speed or the pure residential pool for residential identity and SESSION.
02
Choose billing capacity
Select traffic GB or concurrent request threads.
03
Configure authentication
Use the assigned gateway, port, username, password, or IP allowlist.
04
Monitor requests
Track response time, success rate, retries, and in-flight requests.
Proxy service architecture123Proxy network
ApplicationCLIENT
AuthenticationAUTH
123ProxyROUTE
Public targetRESPONSE
PoolsMixed / pure residential
LocationGlobal random
BillingTraffic / threads
05 / Specifications
Product boundaries and technical facts
Available locations, inventory, bandwidth, and package options are subject to the current console and pricing page.
- Proxy pools
- Crawler mixed pool and pure residential pool
- Crawler pool
- About 95% residential IPs and 5% datacenter IPs
- Pure residential pool
- Residential-only exits with SESSION support
- Exit location
- Global random by default; optional broad presets at extraction
- Billing
- Traffic packages or concurrent thread packages
- Protocols
- HTTP(S) / SOCKS
06 / Developer access
Use one fixed gateway instead of maintaining IP lists
Configure the assigned proxy endpoint in requests, Scrapy, cURL, or another standard HTTP client. The network handles pool selection, health checks, and exit rotation.
Standard HTTP proxy configurationExplicit timeout and retry policyProduction monitoring by workload
Python
import requests
proxy = "http://customer-session-demo:YOUR_PASSWORD@proxy.123proxy.cn:36923"
proxies = {"http": proxy, "https": proxy}
response = requests.get(
"https://target.example/data",
proxies=proxies,
timeout=20,
)
print(response.status_code)07 / FAQ
Questions to answer before production
Validate the product boundary, package model, target behavior, and capacity with representative requests.
What is the difference between the two pools?
The crawler pool mixes about 5% datacenter IPs into a residential-majority pool and is usually faster. The pure residential pool contains only residential IPs and supports SESSION.
Which locations can I select?
Exits are globally random by default. When extracting proxies, you may choose Europe and America, North America, Europe, Asia, United States, or Japan and Korea. Country-level and city-level targeting are not supported.
Which pool supports SESSION?
SESSION is available on the pure residential pool.
How do traffic and concurrency plans differ?
Traffic plans meter transferred GB. Concurrency plans meter in-flight requests and provide unmetered traffic within the selected thread capacity.
Do I need to maintain a local proxy list?
No. Applications connect to a fixed gateway while the proxy network selects and rotates exits.
Test both pools with representative requests
Compare success rate, response time, response size, and concurrency before selecting a production package.