Cloudflare Containers Supercharge Browser Run: Faster, More Scalable, and Ready for AI
A New Chapter for Browser Automation
Browser Run, Cloudflare's platform for programmatic headless browser control, has undergone a major transformation. By rebuilding on Cloudflare Containers, the service now delivers 4x higher concurrency, over 50% faster response times, and significantly improved reliability—all without requiring any changes from existing users. These enhancements are live today, and they open the door to even more rapid feature development going forward.

What Is Browser Run?
Browser Run lets developers spin up and interact with headless browser instances running on Cloudflare's global network. Common use cases include:
- End-to-end testing of web applications in real browser environments
- Investigating suspicious URLs in a secure, sandboxed manner
- Rendering PDF documents and capturing screenshots
- Extracting web content for data analysis or monitoring
- Empowering AI agents to browse and interact with web pages autonomously—a rapidly growing application
With the move to Containers, Browser Run is positioned to be the go-to platform for responsible, large-scale automated browsing.
The Challenge: Outgrowing Shared Infrastructure
Previously, Browser Run shared infrastructure with Browser Isolation (BISO). While technically similar, this arrangement created several bottlenecks:
- Slow startups due to BISO's larger container images, which slowed development and increased cold-start times.
- Suboptimal global distribution—BISO browsers weren't always deployed in the most latency-optimal regions for Browser Run's short-lived sessions.
- Mismatched usage patterns: BISO users maintain long, steady sessions, whereas Browser Run sees short, spiky bursts. This conflict led to scaling delays and availability issues.
To overcome these limitations, the team turned to Cloudflare's own Durable Object–enabled Containers, which entered open beta last year. Adopting this internal platform allowed them to identify and fix pain points before any external customers might encounter them—a prime example of "eating our own dogfood."
The Migration to Cloudflare Containers
The transition was carefully staged to ensure stability and zero disruption for users.
Gradual Rollout with Dual Support
A Worker was inserted into the request path to route a subset of users to Container-powered browsers while others continued using the BISO infrastructure. This dual support enabled direct performance comparisons, isolated implementation bugs, and built confidence in the new approach.

Once validated, the team ramped up adoption in phases:
- First: All Quick Actions endpoints were switched to Containers.
- Next: Workers binding connections for free accounts migrated over.
- Then: Pay-as-you-go accounts were transitioned to test wider stability.
- Finally: All remaining contract customers were moved—with no code changes or worker redeployments needed on their part.
This phased approach kept the service running smoothly throughout.
Performance and Scalability Gains
The results speak for themselves. With the new Container-based architecture:
- Spin-up rate: Developers can now launch 60 browsers per minute via the Workers binding—an 4× increase over the previous limit.
- Concurrency: Up to 120 simultaneous browser sessions are now supported.
- Response times: Quick Action latency dropped by more than 50%, making interactions snappier than ever.
These improvements are live today; existing integrations automatically benefit without any extra configuration.
Faster Feature Development Ahead
Beyond the immediate speed and capacity gains, the move to Containers streamlines internal development. Smaller container images, optimized global distribution, and alignment with Browser Run's usage patterns mean that fixes and new features can be shipped much faster than before. As AI agents increasingly rely on browser automation, Browser Run is now better equipped to scale with demand while maintaining security and performance.
For more details on the technical migration and raw data, refer to the original announcement. No other changes are required—just start experiencing the speed.
Related Articles
- 5 Key Ways Custom Catalogs and Profiles Are Revolutionizing Enterprise MCP Adoption
- Cloudflare Restructures Workforce for an AI-Driven Future
- Cloudflare Unveils Dynamic Workflows: Durable Execution Meets Multi-Tenant Code at Runtime
- How to Implement Enterprise AI with SAP on Azure: A Step-by-Step Guide from Sapphire 2026
- AWS & Anthropic Join Forces on Custom Chips; Meta Commits to Graviton for Agentic AI
- Amazon Redshift RG Instances: Graviton-Powered Performance for Data Warehousing and Lake Queries
- 10 Critical Lessons from the .de DNSSEC Outage Every Administrator Should Know
- Protecting Your Software Supply Chain: A Step-by-Step Security Blueprint for Engineering Teams