Selenium Proxy Setup

How Selenium handles proxies

Selenium itself does not move the traffic — it instructs a browser to do so. That means proxy configuration happens through the browser's options or capabilities rather than through Selenium's core API. Chrome and Firefox each expose their own way of accepting a proxy, and the approach differs slightly between an unauthenticated proxy and one that needs a username and password.

Before writing any code, decide what kind of proxy you need. Datacenter proxies are fast and cheap and suit high-volume testing; residential proxies look like ordinary users and suit tasks where blending in matters; mobile proxies are useful for the most sensitive targets. Match the type to the job rather than defaulting to the cheapest.

Configuring a proxy in Chrome

For a simple host:port proxy in Chrome, you pass it as a command-line argument through ChromeOptions:

from selenium import webdriver

options = webdriver.ChromeOptions()
options.add_argument("--proxy-server=http://123.45.67.89:8080")

driver = webdriver.Chrome(options=options)
driver.get("https://httpbin.org/ip")
print(driver.page_source)
driver.quit()

This works cleanly for proxies that do not require a login. Visiting an IP-echo page like the one above confirms the request is leaving from the proxy rather than your own connection.

Handling authenticated proxies

Most paid proxies require a username and password, and Chrome's --proxy-server flag cannot pass credentials directly. The common solution is a small in-memory extension that supplies the credentials, or a helper library such as selenium-wire that handles auth for you:

from seleniumwire import webdriver  # pip install selenium-wire

options = {
    "proxy": {
        "http": "http://user:pass@123.45.67.89:8080",
        "https": "http://user:pass@123.45.67.89:8080",
        "no_proxy": "localhost,127.0.0.1",
    }
}

driver = webdriver.Chrome(seleniumwire_options=options)
driver.get("https://httpbin.org/ip")

selenium-wire is popular because it keeps authentication tidy and also lets you inspect requests, which is handy when debugging why a proxy is or isn't being used.

Configuring a proxy in Firefox

Firefox uses profile preferences rather than a command-line flag. You set the proxy type to manual and provide the host and port:

from selenium import webdriver

options = webdriver.FirefoxOptions()
options.set_preference("network.proxy.type", 1)
options.set_preference("network.proxy.http", "123.45.67.89")
options.set_preference("network.proxy.http_port", 8080)
options.set_preference("network.proxy.ssl", "123.45.67.89")
options.set_preference("network.proxy.ssl_port", 8080)

driver = webdriver.Firefox(options=options)

For authenticated Firefox proxies, the same extension or selenium-wire patterns apply, since the native preferences do not carry credentials cleanly either.

Rotating proxies in Selenium

If you need many IPs, you have two broad choices. The first is a rotating gateway endpoint from your provider: you connect to a single host:port and the provider assigns a different IP per request or per session, so your code never changes. The second is to manage a pool yourself and start a fresh driver with a different proxy when you want to switch.

Tips for stable rotation

• Prefer a provider gateway when you can — it keeps your code simple and offloads pool management.
• When rotating per session, fully quit and recreate the driver so old cookies and state don't leak across IPs.
• Add retries and timeouts; some proxy IPs will be slower or temporarily unreachable.
• Verify the active IP at the start of a session so failures are easy to diagnose.

Choosing a proxy for browser automation

Selenium spins up real browsers, which is resource-heavy, so a slow or unreliable proxy compounds the cost in wasted time. When comparing providers, weigh connection stability, the freshness and reputation of the IP pool, whether sticky sessions are available, location coverage, and how authentication is handled. Balance those against price rather than chasing the lowest number.

For automation projects on a budget, Cheapest Proxies (our featured value pick) is a strong value-focused option worth considering when you want reliable IPs without overspending — just confirm the proxy type and session options fit your script's needs.

IP leaks that survive a "passing" IP-echo test

The base guide shows how to confirm your proxy with an IP-echo page, but that check is necessary, not sufficient. A browser can route page traffic through the proxy while still exposing your real address through two side channels. WebRTC can reveal local and public IPs directly to a page through STUN requests, bypassing the proxy entirely. DNS resolution can also leak if lookups go through your system resolver instead of the proxy. For any project where your real location matters, disable WebRTC (via browser preferences or a flag) and confirm DNS is handled remotely, then test against a tool that checks for WebRTC leaks rather than just echoing the connecting IP.

Quick leak-hardening steps

• Disable or restrict WebRTC in the browser profile so STUN cannot expose local IPs.
• Prefer SOCKS5 with remote DNS, or confirm your HTTP proxy resolves hostnames remotely.
• Test against a dedicated leak-check page, not only an IP-echo endpoint.

Why a working proxy still gets blocked

A frequent source of confusion is a clean, fast proxy that still hits walls. The cause is usually automation fingerprinting, not the IP. Default Selenium sessions advertise themselves through the navigator.webdriver flag, predictable window sizes, missing or odd headers, and the absence of human-like timing. Sites combine these signals with the IP reputation, so a perfect proxy paired with an obvious bot signature still trips defences. Reducing automation tells — realistic viewport, sensible user-agent, human-paced interactions, and avoiding the most detectable headless markers — often matters as much as the proxy itself.

Architecture: rotate by process, not mid-session

The base guide covers gateway rotation versus self-managed pools. A reliability pattern worth adding is treating each proxy as belonging to its own driver process. Trying to change a proxy on a live driver is awkward and leaves residual state; instead, run one driver per proxy and treat a dead or blocked IP as a reason to tear down and respawn. This makes concurrency cleaner too — you can run several drivers in parallel, each pinned to a distinct IP, and quarantine any that start failing. It also keeps cookies and storage neatly partitioned per IP, which prevents the cross-contamination the base guide warns about.

Choosing proxies for the way Selenium actually fails

Because each Selenium session is heavy, the proxy attributes that matter most are stability under sustained connections and a clean IP reputation rather than raw peak speed. Sticky sessions help when a task spans multiple page loads that must come from the same IP, while a rotating gateway suits broad, stateless collection. Confirm the provider supports the auth method your stack uses and offers the locations your targets expect. For budget-conscious automation, Cheapest Proxies is worth considering as a value pick — verify it offers the session type (sticky vs rotating) and protocol your scripts rely on before scaling up.

Keep comparing with these closely related pages.