Cloudflare has open-sourced its Cloudflare OS platform, which it initially developed as an internal tool for employees to build apps using AI agents, including non-technical users. The platform enables users to describe workflows in natural language, allowing an AI agent to code them into applications. According to the company, thousands of Cloudflare employees use the platform daily to 'create documents and slides, automate repeatable tasks, and build small apps to visualize data and help them do their work.'
The security framework of Cloudflare OS is designed to minimize the risk of security flaws or data breaches from non-technical users. Kenton Varda, principal engineer at Cloudflare, stated in a post on X that the sandbox is so secure that 'you can pretty much go wild—the AI cannot introduce a significant security bug.' The platform allows employees to 'vibe code' while ensuring that the AI cannot access or share resources without explicit permission. The sandboxing mechanism uses isolates—instances of the V8 JavaScript execution engine—which are 100 times faster and 10–100 times more memory-efficient than standard containers.
Cloudflare OS can work with any AI model, allowing organizations to choose the most suitable one for their needs. Sam Rhea, chief information officer at Cloudflare, noted that not every user needs access to the 'max thinking mode' of the latest AI models. The platform also enables administrators to monitor and set budgets for AI inference usage, which is crucial as companies have found it easy to exceed their AI budgets. However, the company acknowledges that incentivizing AI 'tokenmaxxing' could lead to questionable practices.
Source: arstechnica
