OpenClaw Gateway and Proxy: Traffic and Browser Proxy

OpenClaw↗ has two places where “proxy” can appear: (1) Gateway-level — e.g. trusted proxy for incoming requests (load balancer, reverse proxy). (2) Browser-level — the proxy used when an agent drives a browser for scraping or browsing. For residential proxies for web scraping, you care about (2). This guide clarifies the difference and how to set the browser proxy for your agents.

Gateway vs Browser Proxy

Gateway — The OpenClaw Gateway is the process that receives messages from WhatsApp, Telegram, Discord, etc., and routes them to agents. It may sit behind a reverse proxy (e.g. Nginx, Cloudflare Tunnel) for HTTPS or auth. OpenClaw docs mention trusted proxy auth↗ for that. This is not the proxy used for outbound scraping.
Browser proxy — When an agent uses Playwright (or similar) to open web pages, that browser can be configured with an outbound proxy (e.g. residential proxy gateway). That’s where you set Bytesflows or any other residential proxy so the agent’s web traffic goes through rotating residential IPs.

So: Gateway proxy = incoming traffic to OpenClaw. Browser proxy = outgoing traffic from the agent’s browser. For scraping and avoiding blocks, you configure the browser proxy. Why OpenClaw Agents Need Residential Proxies and OpenClaw Proxy Setup.

Where to Configure the Browser Proxy

The proxy is set where the browser is launched — in the skill or agent code that calls Playwright’s launch(), not in the Gateway config file. Add:

javascript

proxy: {
  server: 'http://p1.bytesflows.com:8001',
  username: 'user',
  password: 'pass'
}

to the launch options. Full steps: OpenClaw Proxy Setup and OpenClaw Playwright Proxy. Playwright Proxy Configuration Guide has more examples.

Running the Gateway Behind a Reverse Proxy

If you run OpenClaw on a VPS and put Nginx (or similar) in front for SSL or routing, that’s the Gateway side. The Gateway may need to trust X-Forwarded-* headers; see OpenClaw’s trusted proxy↗ docs. That setup does not change the IP that your agent’s browser uses for outbound requests. For outbound scraping traffic, you still configure the browser proxy (residential proxy) in the Playwright launch options. OpenClaw Browser Automation with Proxies.

FAQ

Where do I set the proxy for scraping? In the browser launch options (Playwright) used by your OpenClaw skill, not in the Gateway config. Add proxy: { server, username, password } to chromium.launch(). OpenClaw Proxy Setup and OpenClaw Playwright Proxy.

Does the Gateway need a proxy? Only if you put the Gateway behind a reverse proxy (e.g. Nginx) for HTTPS or auth; then you may need to configure trusted proxy headers. That’s separate from outbound browser traffic. OpenClaw docs↗.

One proxy for all agents? Yes. Point every browser-based skill to the same residential proxy gateway; the provider’s rotation will assign different IPs. OpenClaw Multi-Agent and Proxy and OpenClaw Rotating Proxy.

Key takeaways

Gateway proxy = incoming (reverse proxy, trusted headers); browser proxy = outbound scraping traffic. Don’t confuse them. OpenClaw Proxy Setup.
Scraping proxy goes in the browser launch (Playwright): proxy: { server, username, password }. OpenClaw Playwright Proxy.
One residential gateway can serve all agents; provider handles rotation. OpenClaw Rotating Proxy and Residential Proxies.
Validate with Proxy Checker and Scraping Test.

Before you start

Gateway = incoming traffic (reverse proxy, trusted headers); browser proxy = outbound scraping. Don’t mix. OpenClaw Proxy Setup.
Residential proxy gateway URL and credentials; add to Playwright launch in your OpenClaw skill. OpenClaw Playwright Proxy.
One gateway can serve all agents; provider rotates IPs. OpenClaw Rotating Proxy and Residential Proxies.
Validate with Proxy Checker and Scraping Test.

When to use this guide

Use this when you need to distinguish Gateway proxy (incoming, reverse proxy) from browser proxy (outbound scraping). Set the scraping proxy in the browser (Playwright) used by your OpenClaw skill, not in the Gateway. OpenClaw Proxy Setup and OpenClaw Playwright Proxy.

Quick tip: One residential proxy gateway can serve all agents; the provider rotates IPs. Validate with Proxy Checker and Scraping Test. Residential Proxies. For Playwright config, OpenClaw Playwright Proxy. See also OpenClaw Multi-Agent and Proxy, OpenClaw Rotating Proxy. For trusted proxy headers (incoming), see OpenClaw docs↗. The browser proxy is set per skill or agent in the Playwright launch options. Use env vars for proxy credentials. OpenClaw Proxy Setup.

Summary

OpenClaw Gateway can sit behind a reverse proxy for incoming traffic; that’s separate from outbound browser traffic. For web scraping and browsing, set a residential proxy on the browser (Playwright launch options) used by your OpenClaw skill. See OpenClaw Proxy Setup, OpenClaw Playwright Proxy, and Residential Proxies.

OpenClaw Gateway and Proxy: Traffic and Browser Proxy

Key Takeaways