Cloudflare is opening up the AI workspace it built for its own employees. Now, partners can use the same foundation to create enterprise agents, automate workflows, and build small internal apps.
The platform connects AI tools with an organization’s internal knowledge, applications, and security controls.
Employees can use the browser-based workspace to research information, create documents, automate tasks, and build what Cloudflare calls Gadgets, or small applications designed around specific business needs.
For systems integrators and managed service providers, the launch creates opportunities around customization, deployment, governance, and ongoing support.
Cloudflare OS combines agents and internal apps
Cloudflare said thousands of its employees already use the platform across engineering, sales, and other business functions.
Users can ask an AI agent to analyze information or complete a task, then turn the output into a working application. These Gadgets can include databases, real-time collaboration, and access controls, allowing teams to create tools without waiting for a traditional software development cycle.
The platform also supports multiple AI providers and self-hosted models. Organizations are not required to standardize on a single model vendor, which could help partners design deployments around a customer’s existing infrastructure, data requirements, and cost priorities.
Cloudflare OS is built on Cloudflare Workers. Its GitHub documentation also says organizations can run the platform on their own infrastructure through the open-source Workers runtime.
Security and integration create channel opportunities
Cloudflare OS uses Cloudflare Access to authenticate users and control requests. Agents begin without permissions and receive access only to the resources required for a specific task.
The platform also includes Gatekeepers, governed connectors that control how agents and applications communicate with internal or external systems. Gatekeepers can restrict access, record agent activity, and require human approval before selected actions are completed.
“Our employees get AI tools connected to the systems they actually use, and security isn’t something IT has to bolt on afterward,” said Matthew Prince, co-founder and CEO of Cloudflare, in the company’s press release.
For channel partners, those controls could support services such as identity configuration, application integration, compliance mapping, agent monitoring, and approval-workflow design.
Cloudflare identified Presidio and Happy Cog as initial implementation partners. The company said systems integrators can customize the workspace for industry requirements, existing technology stacks, and customer-specific compliance needs.
Partners must account for early-access limitations
Cloudflare OS is available through a public GitHub repository under the Apache 2.0 license. Cloudflare also plans to offer a managed deployment through its dashboard, although it has not announced a release date.
The repository describes the August 2026 release as a complete rewrite that remains under heavy development. Cloudflare’s documentation says the current version is “very capable, but still has many rough edges” and labels it an early-access release.
Partners should make that maturity level clear to customers. Early deployments may require additional testing, architecture work, custom integrations, security reviews, and ongoing maintenance before the platform is used for sensitive or business-critical workflows.
Cloudflare OS gives partners more flexibility than a closed AI assistant, but open source does not remove the work required to make enterprise AI dependable.
The strongest near-term opportunity may belong to partners that can combine the platform’s agents and app-building capabilities with governance, integration expertise, and managed services.
See how Cloudflare is equipping AI agents with identities and programmable wallets, opening new integration, security, and payment opportunities for channel partners.