The news
Cloudflare (NET) on October 1, 2026, opened a waitlist for fully managed deployments of Cloudflare OS, a browser-based workspace where employees use AI agents to work with company data, write documents and automate routine tasks. Until now, companies had to deploy and run the open-source software themselves.
Cloudflare first released Cloudflare OS as open source on August 5, 2026, under the Apache 2.0 license, according to the company and Techstrong.ai. Cloudflare said at the time that thousands of people across every function inside the company use it daily. In the October 1 post, Cloudflare said thousands of organizations have started using the open-source version since launch, a figure that has not been independently verified.
The managed version shifts operating work to Cloudflare. In the self-hosted model, Cloudflare noted, someone has to configure the deployment, operate it and keep it up to date. With the managed option, customers decide who can access Cloudflare OS, which organizational skills and context are available, and which systems it can reach, and Cloudflare handles the rest. Customers can also set a custom domain, apply Cloudflare Access policies and choose which AI Gateway to connect.
Cloudflare also described new features. Agents can now connect to Git repositories for code work. Google Workspace integration covers Gmail, Drive, Sheets and Calendar. Agents can create documents, presentations and spreadsheets and export them as Excel, CSV, PDF, Markdown or HTML files, with Word and PowerPoint export listed as coming soon.
Cloudflare did not give a price, a launch date or service-level commitments for the managed service in its announcement. Organizations interested in it can join the waitlist.
The numbers
- Managed waitlist opened
- October 1, 2026
- Open-source release
- August 5, 2026
- License
- Apache 2.0
- Managed pricing
- Not disclosed
Why CEOs should care
For CIOs and IT leaders, the question is no longer whether employees will use AI agents but where those agents run and what they can touch. Cloudflare's design, as described in its August 5 post, starts agents with zero access and routes requests to outside services through gatekeepers that hold credentials and enforce policy. Ask any agent vendor the same questions: what does an agent reach by default, who holds the credentials, and can you see a log of every action?
For CISOs, a managed workspace tied to an existing identity system such as Cloudflare Access may be easier to govern than staff signing up for consumer AI tools on their own. But a managed service you cannot inspect is also a dependency. Before joining the waitlist, ask Cloudflare where data is stored, how long it is kept, and what happens to company skills and documents if you leave.
For CFOs, the missing price is the main open item. Agent workloads can be billed per seat, per action or by model usage, and Cloudflare's model choice runs through AI Gateway. Budget owners should ask how model costs will be passed through and whether they can cap spending by team.
The bigger picture
Cloudflare is entering a crowded field. Techstrong.ai, reporting on the August release, listed Microsoft Copilot for Microsoft 365, Google's Workspace tools with Gemini and UiPath's agent automation among comparable offerings. Cloudflare CEO Matthew Prince said in August that Cloudflare OS is how the company runs itself. Cloudflare's angle is security built on its existing network and identity products, plus an open-source core that companies can still host themselves if they prefer.
What’s next
Watch for Cloudflare to publish pricing, a launch date and service-level terms for the managed version, and for named customers moving from the open-source release to the hosted one.
What “Fact-checked” means
Fact-checking means testing a story’s facts against the evidence before it is published. This story went through at least two separate checks before this version was published.
- What we checked
- Its names, figures, dates, job titles, quotes and who said what were checked against the story’s sources, including its main source where it could be opened. The headline was checked for accuracy and overstatement.
- How
- A first check reviewed the whole story. If it passed, a second, skeptical check went back to the sources to look for mistakes in the most important facts. If a check flagged the story, it was edited to fix the problems found, and a separate re-check then reviewed the whole story again.
- Who
- The checks are made by our newsroom, as steps kept separate from the writing, under rules set by our editor, Hussein Mukhtar. A story the checks still flag is held for the editor, who decides whether it is fixed, published or dropped.
- If something is wrong
- “Fact-checked” does not mean error-free. If a material error is found after publication, we correct the story and add a note saying what changed. Report an error
Companies in this story
