Browser Run Upgraded: Cloudflare Containers Deliver Speed and Scale

Browser Run, Cloudflare's headless browser automation service, has undergone a major transformation. By rebuilding on Cloudflare's own Container platform, we've unlocked higher usage limits, faster performance, and better reliability. You can now spin up 60 browsers per minute via Workers binding and run up to 120 concurrently—4x the previous limit. Quick Action response times have dropped by over 50%. All these improvements are live now with no action required from you. Below, we answer common questions about this upgrade.

What is Browser Run and what are its primary use cases?

Browser Run lets developers programmatically control headless browser instances across Cloudflare's global network. This enables end-to-end testing of web apps, secure investigation of suspicious URLs, and rendering PDF documents. It also supports quick actions like taking screenshots and extracting content. Recently, it has become essential for AI agents that need to interact with the web. Browser Run is designed to be a secure, scalable platform for automated browser tasks, handling everything from simple captures to complex workflows.

What specific performance and scalability improvements have been made?

The upgrade brings a 4x increase in concurrent browser capacity—from 30 to 120. You can now spin up 60 browsers per minute via the Workers binding, compared to the previous 15. Quick Action response times have improved by more than 50%. These gains come from moving Browser Run onto Cloudflare Containers, which offer faster startup times and better global distribution. Additionally, new features and fixes are being shipped more quickly because the Container architecture streamlines development and deployment.

Why did Browser Run need to move away from sharing infrastructure with Browser Isolation?

Previously, Browser Run shared infrastructure with Browser Isolation (BISO). While technically similar, BISO's larger container images led to slower startup and development cycles. More critically, BISO browsers lacked optimal global distribution, hurting latency and resiliency. The usage patterns also clashed: BISO users had long, steady sessions, while Browser Run experienced short, spiky bursts. This caused scaling bottlenecks and availability delays. By migrating to dedicated Cloudflare Containers, we eliminated these conflicts and optimized performance for both services.

How was the migration to Cloudflare Containers carried out without disrupting users?

We began by inserting a Worker in the incoming request path to route some traffic to Container-powered browsers alongside the existing BISO ones. This dual-support phase let us compare performance and isolate bugs. We then gradually increased Container usage: first for all Quick Action endpoints, then for Workers browser binding connections on free accounts, followed by pay-as-you-go accounts. Only after validating stability did we roll it out to all contract customers. The entire transition required no action or redeployment from users.

What are the new usage limits and how do they compare to the old ones?

The old limits allowed spinning up 15 browsers per minute and running 30 concurrently. Now you can spin up 60 per minute and run up to 120 at the same time—a 4x increase in both metrics. Quick Action response times have been cut by more than half. These improvements are immediately available to all users without any configuration changes. Developers can leverage these higher limits for more demanding tasks like large-scale web scraping, parallel testing, or handling AI agent workloads that require many concurrent browser sessions.

How does this upgrade benefit AI agents and other advanced use cases?

AI agents often need to interact with websites dynamically—filling forms, clicking buttons, or extracting data. Browser Run's Container-based architecture provides the speed and scalability needed for these interactions. With faster browser spin-up and lower latency, agents can iterate more quickly. The increased concurrent capacity allows multiple agents to run in parallel, ideal for training or large-scale automation. Moreover, the improved reliability ensures agents can complete tasks without interruptions, making Browser Run a robust platform for AI-driven web interaction.

What does the future hold for Browser Run now that it's on Cloudflare Containers?

We can ship fixes and new features faster than ever because the Container environment simplifies development and testing. The architecture also enables better global distribution, so browsers are launched closer to users for lower latency. We plan to continue optimizing performance, adding more quick actions, and expanding API endpoints. For developers, this means a more responsive platform with higher limits and faster iteration cycles. We're committed to making Browser Run the go-to solution for secure, scalable browser automation at massive scale.