Selenium Proxy Setup: Chrome, Firefox, and Sticky Sessions

This Selenium proxy guide shows how to point WebDriver at Proxy Grove for lawful QA and public-page automation. It is not a guide to evade security on third-party accounts. Use Mobile, Residential, or Corporate Unlimited plans with HTTP or SOCKS5.

Pricing: from $2/IP/day Residential/Corporate, $4.50/IP/day Mobile. See pricing.

Chrome options pattern

Pass --proxy-server=http://host:port (or SOCKS5 form) via Chrome options. For authenticated proxies, many teams use a lightweight auth extension or a local forwarder—pick one approach and standardize. Keep credentials in secrets manager. Pin ChromeDriver versions with your fleet images.

Firefox profile pattern

Set network.proxy.type and related preferences in a Firefox profile. Test HTTP and SOCKS5 separately; do not assume a blog snippet matches your geckodriver version. Document the working preference set in your runbook.

Sticky sessions for multi-step QA

Checkout-like flows need one IP. Configure sticky on the Proxy Grove plan, then keep one WebDriver session on that endpoint. Rotating mid-flow creates false bugs that burn engineering time.

Grid and concurrency

Selenium Grid multiplies load. Map each node pool to a proxy label. Cap sessions per sticky IP. Prefer failing fast with clear proxy errors over silent retries that rotate unexpectedly. Health-check exit IPs from worker nodes, not only from your laptop.

When to prefer Playwright

New projects often move faster on Playwright. If you are greenfield, read the Playwright guide first. Stay on Selenium when legacy suites and corporate grids require it. Puppeteer remains relevant for Chromium-only Node services—see Puppeteer proxy guide.

Also read auth patterns and protocol choice.

Extension-based auth caveats

If you use a Chrome extension to inject proxy auth, pin the extension version and review its permissions. Prefer approaches your security team can audit. Document the pattern so new grid nodes do not ship without it.

Headless differences

Some targets alter behavior for headless Chrome. Debug with a single headed worker before blaming Proxy Grove. Keep user-agent and headless policy consistent with your production risk model and robots or terms constraints.

Grid labeling

Name Selenium node pools after proxy labels. When a pool misbehaves, you can drain it without pausing the entire grid. Export node logs with proxy label fields for incident review.

Next steps on Proxy Grove

Compare live Unlimited pricing, open residential, mobile, or corporate, then create endpoints in the app. For protocol choice read SOCKS5 vs HTTP. For meter math read unlimited vs pay-per-GB.