Generic Warm Start
Instant attach to a pre-booted pool browser. Best for general scraping, stateless agents, and fast testing. Happens when your request has no specific identity or storage requirements.
Cloud browsers usually take seconds to boot. BrowserCity gets your agents clicking in milliseconds. Here are our real-world latency numbers—measured from the moment your request hits our edge until the remote browser is primed and ready to navigate.
We keep a massive pool of browsers pre-warmed and ready to go. But depending on how much custom state you need to inject before starting, your startup time will vary. Here is exactly what to expect.
Instant attach to a pre-booted pool browser. Best for general scraping, stateless agents, and fast testing. Happens when your request has no specific identity or storage requirements.
Applying a specific stealth fingerprint or injecting localStorage state before handing you the session.
The worst-case scenario. If a massive traffic spike temporarily outpaces our self-adaptive warming pool, we boot a fresh runtime in about 1.5 seconds.
A benchmark is only useful if it measures the full journey. Here is exactly how we measure our infrastructure performance.
We measure the entire journey. Our metrics capture the time from when your API request arrives to the exact moment the remote browser is started, primed, and fully ready to navigate via MCP or Playwright.
Spinning up a container is fast, but shaping it to match a specific device or injecting state takes time. We explicitly separate generic warm starts from custom ones so you know exactly what your specific payload costs.
Our warm pool uses a self-adaptive algorithm to match supply with demand. But if an unpredictable traffic spike drains the pool, you fall back to a 1.5s cold start—not a queued timeout.
Connect Playwright, your MCP client, or trigger a REST action and feel the sub-250ms difference for yourself.
Start for free. No credit card required. Private sessions by default.