director of eng, developer platform @cloudflare
Joined August 2013
- Tweets538
- Following1.2K
- Followers1.9K
- Likes3.5K
🤸🏽 retweeted
very excited for Cloudflare to be part of this
the best part of this new protocol is that it bets hard on existing standards — OAuth, CIMD, dPoP, MCP, OpenAPI
if you have a remote MCP server today, you should have a clear path towards adopting this
On Tuesday, Sierra announced Personal Agent Protocol with Meta, Genesys, Instinct, Rocket, Shopify, Stripe, and Walmart. The response has been incredible and today we’re publishing the first draft of the protocol at personalagentprotocol.org and announcing 35 new design partners.
As personal agents take on more, our goal is simple: consumers have choice and control, and companies know when they’re dealing with an agent and who it represents.
This draft is a starting point. We’d love for companies, developers, and personal agent builders to read it, give feedback, and help us build an open standard together.
🤸🏽 retweeted
Welcome @rough__sea and the @Deno_Land team to Cloudflare.
Big plans to come!
The team behind @Deno_Land, including @rough__sea (he knows a bit about Node.js), is joining @Cloudflare.
They’ll help make Workers better for Node.js developers and advance workerd as a core primitive for what comes next. cfl.re/4zlzlSa
🤸🏽 retweeted
npm i -g cf
a CLI for the entire Cloudflare API, designed specifically for agents. with Vite as the dev server, and Rolldown as the bundler
cloudflare.config.ts
configure your Worker, and soon all of Cloudflare, in TypeScript. validate config by just running tsgo or tsc
blog.cloudflare.com/cloudfla…
Also launching today, and this is the big one: cf in open beta.
cf is the new CLI for the whole of Cloudflare. Where Wrangler focused on Workers, cf exposes the whole API surface for agents.
Testing this has made using Cloudflare so easy.
blog.cloudflare.com/cloudfla…
Change the API. Codegen breaks. Nobody finds out until release.
We hit this at 3,500+ operations, so we built Forge. Now open source:
• Lints every API change in CI
• Preview SDK, CLI & docs builds per PR
Massive effort from this team that has been a long time in the making.
blog.cloudflare.com/forge-op…
bless
Sometimes it's easier to show than tell. We're sure this update will help with that. 👀
GitHub CLI now has a repeatable --attach flag that uploads a local image or video. Reference it inline in an issue, pull request, or comment body.
Available now to all users on GitHub across all plans. 🎉
github.blog/changelog/2026-0…
🤸🏽 retweeted
This week I "only" produced a 13,000-line PR...
Pretty excited about this feature though! I will now be able to vibe code my ideal code review workflow as a Cloudflare OS gadget.
Specifically, for reviewing PRs sent to me from others. PR descriptions are inconsistent and inevitably don't cover the points I care about, so I want an agent to look first and generate a report tailored for me, so once I open the PR I can get my bearings quickly. Then I want to be able to ask the agent further questions as I read, and maybe even prompt it to make changes, which should just be pushed directly back to the branch. (Sorry, "suggestions" are weak, I just want to push code to your branch, deal with it.)
I know there are many AI-assisted code review tools out there that do similar things, but none of them do things exactly the way I want, nor can I immediately add new features to them as I think of them. That's why I want to be able to vibe code my own -- but safely. This is exactly what Cloudflare OS and personal software is all about...
github.com/cloudflare/cloudf…
🤸🏽 retweeted
Most agents can build apps. Cloudflare OS can actually use the apps it builds.
This means if your team builds a tool to do a job, an agent can use the tool as you would—on a schedule or upon a triggering event—even if you're not there to work with it.
listen to @dillon_mulroy 👇🏾
the security model behind cloudflare os, gatekeepers, and isolates solves this
point your clanker at this repo and ask it to teach you about the security model
github.com/cloudflare/cloudf…
🤸🏽 retweeted
CloudflareOS is everything that you wish Google Workspace could do well, but can't
just went from data question to spreadsheet to pricing memo with a real-time editor in one browser tab
then sent someone on my team a link to all of it (including original prompt) — we can both keep prompting and make direct edits
all of this is possible to get done with codex/claude via computer use and integrations — but i am telling you it is night and day faster this way, and I don't worry about security/permissions at all because cfos makes the footguns impossible, in a way nothing else can
what starts out as a toy...
os.cloudflare.app/
We've now made it even easier to get started with CloudflareOS:
→ You can now add the context and scheduled task gatekeepers through our deploy tool
Use the context gatekeeper to create skills, documents, and other preferences you want your agent to have. Scheduled tasks let your agents wake up later to follow up, run jobs, and keep things moving.
We made getting started with Cloudflare OS ridiculously easy.
One-click deploy to your own Cloudflare account:
→ Sign in
→ We spin up the Workers, KV namespaces, R2 bucket + public URL for you (takes ~1 minute)
Or you can fork the open-source code and make it fully your own. Your own agent workspace, ready to go.
Try it → os.cloudflare.app/deploy
🤸🏽 retweeted
It's been very funny to me to watch rando larpers spend the last day trying to dunk on Kenton with various "OS engineer here! this is not an OS" takes.
I have worked on more OSes than >99.999% of the people on this website. This is an OS. It's not the 80s anymore.
Today we are releasing Cloudflare OS, a chatbot with connectors, just like every other tech company is doing.
Except actually, it's different. This is a remake of Sandstorm[.]io, my startup from 10 years ago, except this time built on Cloudflare Workers (the platform I've spent the last 9 years building) and deeply leveraging AI. This is more or less the culmination of my secret 10-year master plan.
This is a full-on personal app vibe coding platform, in which the sandbox is so secure that you can pretty much go wild -- the AI cannot introduce a significant security bug. We believe a company's security team can feel comfortable giving non-technical users permission to vibe code and then sleep soundly at night.
How is that possible? It's the Sandstorm security model, revisited. A "Gadget" is the same thing as a Sandstorm "Grain": a fine-grained app instance. For example, if you have a document editor app, each document runs as a separate instance of the app, in a separate sandbox (one "Gadget").
This means two things, both of which I think are Big Deals:
1. The platform can manage all access control, by controlling who can access the Gadget at all. There is no way the Gadget can accidentally leak itself to an attacker -- even an attacker who has access to other Gadgets based on the same app.
2. Since everyone is running their own copy of the code, everyone can freely *modify* their copy of the code.
Think about #2 a bit more.
What if, when you wanted a new feature in the software you are using, you could just prompt your agent to add it?
This doesn't work in the cloud Software-as-a-Service model, because you are not running your own copy of the app.
Sandstorm tried to change that 10 years ago, but the world wasn't ready, because not enough people had the skills or patience to actually modify their software. AI has changed that. Now you just ask the agent -- the same agent that you are using to help you interact with the Gadget can also modify the code of the Gadget.
And it is so fun.
github.com/cloudflare/cloudf…
🤸🏽 retweeted
I paired Cloudflare OS with self hosted deepseek-v4-flash and the result is freakin amazing
Imagine running this setup inside DGX spark, you get fully private work companion solving all your work stuff
The thing about this type of software is it already have a good harness and it probably can work out of the box with every open weight model
Cant wait for the future
Today we are releasing Cloudflare OS, a chatbot with connectors, just like every other tech company is doing.
Except actually, it's different. This is a remake of Sandstorm[.]io, my startup from 10 years ago, except this time built on Cloudflare Workers (the platform I've spent the last 9 years building) and deeply leveraging AI. This is more or less the culmination of my secret 10-year master plan.
This is a full-on personal app vibe coding platform, in which the sandbox is so secure that you can pretty much go wild -- the AI cannot introduce a significant security bug. We believe a company's security team can feel comfortable giving non-technical users permission to vibe code and then sleep soundly at night.
How is that possible? It's the Sandstorm security model, revisited. A "Gadget" is the same thing as a Sandstorm "Grain": a fine-grained app instance. For example, if you have a document editor app, each document runs as a separate instance of the app, in a separate sandbox (one "Gadget").
This means two things, both of which I think are Big Deals:
1. The platform can manage all access control, by controlling who can access the Gadget at all. There is no way the Gadget can accidentally leak itself to an attacker -- even an attacker who has access to other Gadgets based on the same app.
2. Since everyone is running their own copy of the code, everyone can freely *modify* their copy of the code.
Think about #2 a bit more.
What if, when you wanted a new feature in the software you are using, you could just prompt your agent to add it?
This doesn't work in the cloud Software-as-a-Service model, because you are not running your own copy of the app.
Sandstorm tried to change that 10 years ago, but the world wasn't ready, because not enough people had the skills or patience to actually modify their software. AI has changed that. Now you just ask the agent -- the same agent that you are using to help you interact with the Gadget can also modify the code of the Gadget.
And it is so fun.
github.com/cloudflare/cloudf…
🤸🏽 retweeted
Cloudflare OS is an open-source platform that lets everyone in your company build apps, automate work, and safely access internal systems, shaped around what your organization knows and how it operates. cfl.re/4w6ViT0
🤸🏽 retweeted
introducing @cloudflare os, an open source ai productivity platform
there are a lot of tools to help developers be more productive with AI, but ramping the rest of the workforce is still a major challenge
you need built-in context, skills AND access to systems of record. handing out thousands of API keys is a recipe for disaster
this is where what we built is truly unique because it has a built-in governance model that makes it easy to delegate existing permissions and narrowly scope them for the particular task
and because it's open-source, you can customize it to fit your orgs needs. the reason everyone seems to be building one of these is: one size fits all doesn't cut it for something that every employee needs to touch
Cloudflare OS is an open-source platform that lets everyone in your company build apps, automate work, and safely access internal systems, shaped around what your organization knows and how it operates. cfl.re/4w6ViT0
🤸🏽 retweeted
This isn't just about agentic work, delegated authority, and vibe coding - The platform we are introducing is about personal software and a radically different approach to the software distribution model that is needed to support it.
Cloudflare OS enables humans to build, using agents, the exact software they want to accomplish exactly what they need. It's about humans using agents to encode determinism into workflows, dashboards, and even custom agents.
Personal software is private, by default - but sharable and multiplayer. Personal software acts on behalf of you, the human, not a service account. Personal software has an interface that is designed for humans AND agents alike.
I truly believe this is the beginning of one of the most ground breaking technologies since Workers themselves were announced. This is Workers for the agentic software era.
We built Cloudflare OS to equip our teams to safely rethink how they get work done with AI. The platform brings together the best of our technologies, from our Compute primitives to our Zero Trust suite. cfl.re/4yTCRUN
🤸🏽 retweeted
Cloudflare OS "the spiritual successor of Sandstorm" is open source today.
One app per document. Capabilities instead of credentials. Nothing reachable that you didn't hand it.
@kentonvarda got that model right with Sandstorm years before the infrastructure existed to run it. Cloudflare OS is that model on Workers, with an agent in the loop.
blog.cloudflare.com/cloudfla…
Today we are releasing Cloudflare OS, a chatbot with connectors, just like every other tech company is doing.
Except actually, it's different. This is a remake of Sandstorm[.]io, my startup from 10 years ago, except this time built on Cloudflare Workers (the platform I've spent the last 9 years building) and deeply leveraging AI. This is more or less the culmination of my secret 10-year master plan.
This is a full-on personal app vibe coding platform, in which the sandbox is so secure that you can pretty much go wild -- the AI cannot introduce a significant security bug. We believe a company's security team can feel comfortable giving non-technical users permission to vibe code and then sleep soundly at night.
How is that possible? It's the Sandstorm security model, revisited. A "Gadget" is the same thing as a Sandstorm "Grain": a fine-grained app instance. For example, if you have a document editor app, each document runs as a separate instance of the app, in a separate sandbox (one "Gadget").
This means two things, both of which I think are Big Deals:
1. The platform can manage all access control, by controlling who can access the Gadget at all. There is no way the Gadget can accidentally leak itself to an attacker -- even an attacker who has access to other Gadgets based on the same app.
2. Since everyone is running their own copy of the code, everyone can freely *modify* their copy of the code.
Think about #2 a bit more.
What if, when you wanted a new feature in the software you are using, you could just prompt your agent to add it?
This doesn't work in the cloud Software-as-a-Service model, because you are not running your own copy of the app.
Sandstorm tried to change that 10 years ago, but the world wasn't ready, because not enough people had the skills or patience to actually modify their software. AI has changed that. Now you just ask the agent -- the same agent that you are using to help you interact with the Gadget can also modify the code of the Gadget.
And it is so fun.
github.com/cloudflare/cloudf…
🤸🏽 retweeted
Today we're open sourcing Cloudflare OS:
- An agent workspace grounded in skills your team curates.
- A platform for personal, modifiable apps that people can build and share.
- A new security framework for safe access to internal data.
You can use it to:
🤸🏽 retweeted
Today we are releasing Cloudflare OS, a chatbot with connectors, just like every other tech company is doing.
Except actually, it's different. This is a remake of Sandstorm[.]io, my startup from 10 years ago, except this time built on Cloudflare Workers (the platform I've spent the last 9 years building) and deeply leveraging AI. This is more or less the culmination of my secret 10-year master plan.
This is a full-on personal app vibe coding platform, in which the sandbox is so secure that you can pretty much go wild -- the AI cannot introduce a significant security bug. We believe a company's security team can feel comfortable giving non-technical users permission to vibe code and then sleep soundly at night.
How is that possible? It's the Sandstorm security model, revisited. A "Gadget" is the same thing as a Sandstorm "Grain": a fine-grained app instance. For example, if you have a document editor app, each document runs as a separate instance of the app, in a separate sandbox (one "Gadget").
This means two things, both of which I think are Big Deals:
1. The platform can manage all access control, by controlling who can access the Gadget at all. There is no way the Gadget can accidentally leak itself to an attacker -- even an attacker who has access to other Gadgets based on the same app.
2. Since everyone is running their own copy of the code, everyone can freely *modify* their copy of the code.
Think about #2 a bit more.
What if, when you wanted a new feature in the software you are using, you could just prompt your agent to add it?
This doesn't work in the cloud Software-as-a-Service model, because you are not running your own copy of the app.
Sandstorm tried to change that 10 years ago, but the world wasn't ready, because not enough people had the skills or patience to actually modify their software. AI has changed that. Now you just ask the agent -- the same agent that you are using to help you interact with the Gadget can also modify the code of the Gadget.
And it is so fun.
github.com/cloudflare/cloudf…