Web scraping
Scrapers that stop
getting blocked.
Most blocks don't come from your proxy. They come from a browser that reports a software GPU,
no plugins and webdriver: true — a combination no real visitor has. Change the machine and the
block goes away.
Built by a team that operates browser fleets in production for clients today.
- brands
- "Google Chrome", "Chromium", "Not)A;Brand"
- platform
- MacIntel
- renderer
- ANGLE Metal Renderer: Apple M2
- headless
- false
- webdriver
- false
- plugins
- 5
Pick the cost per page
Nano at $0.03/hr for cooperative sites, Apple Silicon at $0.40/hr for the ones that read the GPU. Same API, one parameter apart — route each domain to the cheapest class that passes.
Bring your own proxy
Point a session at your residential or mobile pool at no markup. Egress is fail-closed: if the proxy is unreachable the browser doesn't launch, so a run can never leak a datacenter IP.
Scale by the minute
One browser or two hundred, billed only while connected. No capacity to size in advance and no idle fleet to keep warm between crawls.
Logged-in targets
Sessions that stay signed in.
Marketplaces, portals and social sites show more to a logged-in visitor — and treat a fresh login on every crawl as a signal. A persistent profile logs in once and stays in, crawl after crawl, without the re-auth churn that gets accounts flagged.
// 1. create a profile once — it's the customer's logged-in state
const profile = await bf.createProfile("shop-eu");
// 2. run a session on it, log in, do work…
const s1 = await bf.createSession({
image: "chrome-stable", machineClass: "mac-shared",
profile: { mode: "managed", id: profile.id },
});
// …end it. The state is saved automatically because the session ran on the profile.
await bf.endSession(s1.id);
// 3. next week: a new session on the same profile is already logged in
const s2 = await bf.createSession({
image: "chrome-stable", machineClass: "mac-shared",
profile: { mode: "managed", id: profile.id },
});Responsible use
What we expect.
BrowserFleet is for collecting publicly available data and automating things you're entitled to automate: price monitoring, SEO and SERP data, market research, QA of your own properties. It is not for credential stuffing, account takeover, bypassing paywalls or evading a site's stated terms. Accounts doing that lose access. Our responsible-use policy →
Try it on the site that's blocking you.
Five hours of Apple Silicon on the free credit. If it still gets blocked, tell us which site — that's a bug we want.
Built by a team that operates browser fleets in production for clients today.