Guillermo Rauch

An awesome day for humanity ✦ · ✧ · ✦ · ✧ ✧ · ✦ ✧ · ✧

Got to experience the magic of @uisdk. Two prompts of fx + Gemini Argon. Pure Web Components. No framework. The agent reused 14 UI SDK behaviors out of the box 🤯: window resizing, drag, menubar hover-switching, typeahead, roving toolbar focus, 10-band EQ sliders, scoped hotkeys, ARIA, and WebMCP agent control. https://winamp-ui-sdk.vercel.app/ (source linked) Music by @suno.

Subagents on https://fx.sh work so well… extremely delightful. Amazing how much goodness fits in 6.33mb 😁 All my projects are built by planning with the smartest model + delegating to fast model subagents. It's a 1-line instruction in ~/.𝚏𝚡/𝙰𝙶𝙴𝙽𝚃𝚂.𝚖𝚍.

So… what did you get done this week?

Life is majestic

This is how agents will ship all code. → Agent writes feature-gated code → Verification gates on push (proof checking, security audits) → Agent monitors metrics You can live the future today with Vercel CLI: 𝚟𝚎𝚛𝚌𝚎𝚕 𝚏𝚕𝚊𝚐𝚜, 𝚟𝚎𝚛𝚌𝚎𝚕 𝚖𝚎𝚝𝚛𝚒𝚌𝚜, 𝚟𝚎𝚛𝚌𝚎𝚕 𝚍𝚎𝚙𝚕𝚘𝚢.

Co-sign

Vercel Sandbox is inevitable

RT Vercel Developers Vercel Security Review no longer requires a Vercel project to scan your GitHub repos. Every push to a connected repo gets code changes reviewed for vulnerabilities ↓ https://vercel.com/changelog/vercel-security-review-no-longer-requires-a-linked-vercel-project

Workflow SDK 5 and Vercel Workflows are a monumental step forward for agents. ◾ The 'time to first step' and 'time between steps' have massively improved. ◾ We doubled down on encryption and security. ◾ The reliability of the overall system has been proven at massive scale. We've ironed out many subtle details and exercised many disaster-recovery scenarios. ◾ We made the underlying durability log reusable for any agent harness. 𝚍𝚞𝚛𝚊𝚋𝚕𝚎-𝚑𝚊𝚛𝚗𝚎𝚜𝚜 was born. Check it out:

RT Chris Tate UI SDK: a library of behaviors for agents to build the next generation of UIs → 50+ behaviors: popups, selection, typeahead, focus → Components for React, Vue, Svelte, Solid, Web Components → Easy for agents to build with, and for any agent to use What agents made with it ↓

What happens to frameworks in the agents era? to component libraries? to @shadcn? The answer is: we need to drill deeper. Delete the part, optimize, accelerate, automate. UI SDK is like the x86 instruction set for agents to build new UI components spontaneously. Agents are the new design engineers, and they're ready to build *anything* you ask, tailored to your usecase. High-level, rigid, single-framework abstractions are like kryptonite for what is essentially an infinite, untiring creation machine. I think we've found someone that allows us to have the cake and eat it too. A decades-long dream. UI SDK is packaged behaviors that give us accessibility, performance, cohesion, saving us tokens, errors and effort, but with absolutely zero limitations and tradeoffs for what UIs we can generate. Enjoy!

PSA: creating wonderful products still requires extraordinary amounts of human focus, effort, motivation, and love. And an attention span.

We have a benchmark that makes 𝚏𝚡 bootup compete with an ultra-optimized pure-assembly 𝚏𝚡.𝚜 version 😂

Beam by Reflection is free on https://fx.sh until Nov 12. American open weights. Zero data retention. Enjoy 🇺🇸

Rust is a superb language and ecosystem, especially given it was invented pre agentic coding. With agents, new languages will blossom, even purpose-built ones for specific usecases, and discretionary C or even LLVM IR, assembly… AI is fluid. No more one-size-fits-all.

Vercel Security Review, less than 1 week since launch: ✓ 𝟱𝟳,𝟴𝟵𝟳 vulnerabilities identified ✓ 𝟲,𝟬𝟮𝟵 critical or high ✓ 𝟭𝟰,𝟲𝟲𝟱 already patched by Agent I'm shocked at ① the number of critical vulnerabilities that would go unnoticed and ② the ease at which the agent automatically fixes them. You can turn it on for any repo, no matter where you deploy to.

So bullish on MCP. Its utility exceeds the "M" part: even AI aside, it'll be a civilization-scale accomplishment to get all our systems communicating through one universal open protocol. A missing piece was how to get arbitrary callbacks across sessions. MCP could read and write, but it couldn't tell you later "this event happened." MCP Events fixes this. We're adding support in our own MCP. Say an agent makes a deployment, and wants to know when it's complete. Instead of polling, it can subscribe to 𝚍𝚎𝚙𝚕𝚘𝚢𝚖𝚎𝚗𝚝.𝚜𝚞𝚌𝚌𝚎𝚎𝚍𝚎𝚍. PSA: With Vercel Queues and Workflows, writing the backend machinery for your own MCPs to do this is easy peasy.

At @vercel we got to test this model before release. Best way to summarize it: astonishingly brilliant. One interesting takeaway: it's very clear that this new generation of models can take on unprecedented bodies of parallel work. To actually seize this intelligence, you need cloud sandboxes. Full stop. The model can delegate and orchestrate far more than a single computer can handle. Likely far more than you can even imagine it can handle. Your bottleneck is now ① your ambition and ② your parallel computing power. Btw, I learned from @cramforce you don't need to overcomplicate it. Install Vercel CLI, setup the Plugin and tell the agent to use /𝚘𝚏𝚏𝚕𝚘𝚊𝚍-𝚝𝚘-𝚜𝚊𝚗𝚍𝚋𝚘𝚡 skill. It'll provision and de-allocate cloud sandboxes as needed. You don't need a "multi-agent remote harness ralph orchestrator" 😆

Built a physical simulation of a record player with WebGL2 and Web Audio. It's *shockingly* accurate: http://vinyl-gl.vercel.app. Instead of playing a song, it works by converting it into the physical shape of a groove. Then a simulated stylus rides those walls 384,000 times a second, and you hear the voltage that motion would induce in a cartridge. So the crackle, the pops, the wow/flutter, the dull inner grooves are natural byproducts of the simulation. You can also "DJ" the record. (The converted songs are in the public domain or CC-licensed. Code is linked for you to remix and re-deploy.)

Imagine we get gta6.wasm before gta6

Somewhere, something incredible is waiting to be known

Generative UI for conversational agents. Combining @aisdk with Chat SDK and 𝚓𝚜𝚘𝚗-𝚛𝚎𝚗𝚍𝚎𝚛, you can go straight from LLM to UI on platforms like Slack and Teams. I believe this will become the dominant Human-Agent Interaction (HAI) paradigm: https://vercel.com/changelog/json-render-and-chat-sdk-agents

I unironically agree with @bcherny’s take that “coding is solved, but bugs are not.” The deeper meaning is that you can in theory code *anything* now. You can burn infinite tokens refining and optimizing. Including optimizations that no one will care about 💀 With Vercel you can build a different kind of recursive improvement loop. You capture traces from production, you then optimize and course-correct based on how your software is being used IRL. This is not a new idea. The best engineers I’ve ever worked with never jumped into coding (now prompting) right away. You instrument the system, collect data, understand distributions, user journeys, and then put fingers on keyboard. You can ask your agent today to look at your traces!

① Get an X sub ② Build with Bot / Cursor ③ Deploy to Vercel, scales from 0 to 🌙 ④ Launch and distribute on X ⑤ Monetize with X Money / Creators The loop is nearly complete ♾️

RT Vercel Deploy apps in one click with @cursor_ai. • No account to create • No deployment configuration • Database provisioning included Built on Vercel for Platforms → https://vercel.com/for/platforms

SpaceXSI 🤝 Vercel Grok and Cursor can now build, deploy, and scale https://vercel.com/for/platforms

“Doing stuff for you over time” is the killer app of agents. Muse and Instinct consolidate memories at night, remind you of stuff in the morning, message you after N days passed since your last interaction… Vercel Schedules are like Cron on steroids. They can be defined at runtime, scale infinitely, and sit outside of the Sandbox, which means you can cost-efficiently invoke compute as needed.

I had a hard time believing this… but I had https://fx.sh build a repro with 𝚊𝚕𝚙𝚒𝚗𝚎:𝟹.𝟸𝟺. ① Type in a wrong password ② Type in a > 256 chars password ③ root Alpine team has pushed a fixed apk for 𝚘𝚙𝚎𝚗𝚜𝚜𝚑-𝚜𝚎𝚛𝚟𝚎𝚛 (𝟷𝟶.𝟹_𝚙𝟷-𝚛𝟸)

"you will own nothing and be happy" "you will know nothing and be happy" ^ Gotta prevent this at all costs. Use AI to know and understand.

We will all be learning in coming weeks & months, that most human-made software: ▫️ Was really slow ▫️ Had many security vulnerabilities ▫️ Was under-specified and incorrect AI will rewrite it all. Many times. It will make it fast, cost-efficient, and secure. It will think about failure modes we couldn't wrap our heads around. It will prove, formally verify, load test, harden. It will turn memory-unsafe foundations into robust, optimized ones. If you ask, it'll thoroughly document it, teach you about it, make diagrams, animations, and slide decks. In many ways, we are witnessing the true birth of software engineering.

In just 24 hours 🤯 ✓ 𝟲,𝟴𝟭𝟳 vulnerabilities identified ✓ 𝟱𝟴𝟬 critical or high ✓ 𝟭,𝟰𝟭𝟭 already patched by Agent Vercel Security Reviews are 100% free until Nov 10

It’s so inspiring to work with people who care so much about building quality products. Greatness is the byproduct of a million iterations. Agents or otherwise.

`𝚟𝚎𝚛𝚌𝚎𝚕 𝚍𝚘𝚌𝚝𝚘𝚛` - recommend you ship doctor subcommands in your CLIs!

Vercel CI accelerates any repo Vercel Security Review protects them

RT Vercel Developers Vercel Security Reviews are free until November 10.

I've been looking forward to this day, because Vercel has an internal security agent that is **astonishingly effective** (lovingly called @𝚑𝚊𝚌𝚔𝚎𝚛). Every single line of code gets audited, automatically. You can now get the same power. Just enable it on all your repos (this product works whether you host on Vercel or not) and your peace of mind will increase tremendously. You'll see it catch things that every other agent misses. Why: you cannot just bank on your engineers having a skill installed, prompting "please make secure", or rolling the weights' dice that the code is good out of the gate. The other cool thing about this is that this was entirely developed by our security team. A lot of companies out there are probably in a mad rush right now trying to staff up. If you wanted to 'download a security team' as software, this is the closest you get today.

npm i -g ai-cli Such a simple idea. An 𝚊𝚒 CLI. Gives every agent access to video, voice, image, text generation. And 𝚊𝚒 𝚍𝚎𝚌𝚒𝚍𝚎 brings Jev-class models. All behind AI Gateway with cost and ZDR controls.

So excited about this. We're rolling out a mechanism that protects Vercel customers even in a Binance-style attack, where your engineers get targeted with infostealer malware. With DBSC, renewing a session requires a key bound to your device, which means stolen cookies can't be practically re-used from remote Command & Control (C2) infrastructure. While this is a pretty major win, there's no silver bullet in security. Make sure your Vercel account has MFA enabled (both app-based TOTP and passkeys), your organization enables SAML / directory sync, your endpoints are protected, etc. (and stay paranoid!)

I just got this gift from a customer well played

So bullish on WebMCP

So cool!

Stateless sandboxes are the wrong benchmark. Agents are only useful if they have their workspace with them (identity, skills, repos, memories…) Vercel Sandbox team has been beautifully delivering on their latency & throughput SLOs, while prioritizing reliability and security. Doing this while shipping clusters in 19 (‼️) global regions, new features every week, and under immense amount of customer demand in the double digit growth percentage per week.

Securing the supply chain is one of our top priorities. When npm packages get hijacked, our system blocks their deployment, protecting your code and secrets from exfiltration.

RT Malte Ubl Given the deserved attention this got we posted a security bulletin https://vercel.com/kb/bulletin/linux-kernel-security-advisory-cve-2026-98432 TL;dr: Vercel has completed the application of Linux kernel security updates to address CVE-2026-98432.

F1 is basically the perfect competition for our times. The celebration of extraordinary technology and extraordinary human ability. That’s what Vercel stands for! (also: winning)

Own your intelligence. Anyone that opposes this has something to sell you (perhaps your own intelligence back to you.) Local models need to continue proliferating. The “absolute unit” massive models may be out of reach for now locally, so your best option is to, indeed, keep your options open. Own your harness, data, connections, runtime: your agents. Embrace open source.

Working with Gemini on https://fx.sh is hilarious 😂😂 Other than being SUPER EXCITED!, I've been testing it on two high complexity projects and it's doing great.

A society fit of body, mind and soul is a winning society

Turns out alchemical dreams were real. We converted sand into superintelligence. Who cares about gold… We transmuted the most ordinary into the most extraordinary.

This would go so hard in SF

We're in humanity's golden age

An agent on every desktop

RT Go 👀

The rate of improvement on https://fx.sh is bewildering. It's now even smaller, faster and more token-efficient. We're continuously benchmarking and asking it to improve its own performance across open and proprietary models.

Next Native is coming. The whole world will be natively optimized in the coming months. Safe, performant, and cost-efficient.

Vercelnetes or V8S for short

Enjoy!

Amazing technology is being developed and released at a rate humans can hardly keep up with. Agents are shipping software faster than we can yap about it.

I love how quickly our team works to bring the latest LTS and security upgrades for @nodejs. Should be a nice performance boost given 𝚏𝚎𝚝𝚌𝚑 now uses HTTP/2 by default, and agents produce a lot of network traffic

▲ 👻 🎃 🕷️ https://vercel.com/home

The future of software engineering is agents talking to other agents. Even across different companies. We’re learning daily from the feedback agents send us on how to make a better Vercel. Put in your 𝙰𝙶𝙴𝙽𝚃𝚂.𝚖𝚍: send Vercel feedback when something could be better.

This is like DNS for models. An alias that *resolves* to a model, giving you traffic-steering abilities. Solves a huge number of problems. From not having teams stuck on old model versions, to saving on costs, to removing “decision paralysis”. Try it!

Excited to partner with @suhail to serve Particle's models and inference via @vercel AI Gateway. Particle's mission is to make intelligence as accessible and abundant as electricity. Together, we'll make this a reality for every developer and enterprise.

Vercel CI solves the "AI coding bottleneck" via: ① Fast and stable infrastructure (faster work) ② Turborepo integration (less work) ③ Agentic CI ① is obvious. Your CI is slow, flaky, goes down. We solve that. We operate 20 worldwide clusters doing millions of microVMs per day. ② is less obvious, but very powerful. Turborepo was designed to intelligently skip and cache prior work. The graph of dependencies is computed, not written. Ask your agent to migrate to @turborepo and start saving money and time. Even if you don't use us. But you should, it's really good. ③ is a game-changer. 𝚟𝚎𝚛𝚌𝚎𝚕 𝚌𝚒 in the CLI is the first step, because now your agents can inspect, abort, trigger, replay CI. Vercel Agent can help optimize it. And this is just the start. You can expect a lot of hooks and capabilities to make CI something far more intelligent, that can even capture the nuance and judgment an engineer would have ("Are those tests even worth running?") Agentic CI is the future of how we ship software. Excited to build it with you all!

Theo cooked

fx is now much faster, smaller and efficient Try https://fx.sh on your iPhone Duo

$ 𝚊𝚕𝚒𝚊𝚜 𝚗𝚘𝚠=𝚟𝚎𝚛𝚌𝚎𝚕 $ 𝚌𝚊𝚝 >𝙳𝚘𝚌𝚔𝚎𝚛𝚏𝚒𝚕𝚎 <<'𝙴𝙾𝙵' 𝙵𝚁𝙾𝙼 𝚘𝚟𝚎𝚗/𝚋𝚞𝚗:𝚊𝚕𝚙𝚒𝚗𝚎 𝙲𝙼𝙳 ["𝚋𝚞𝚗","-𝚎","𝙱𝚞𝚗.𝚜𝚎𝚛𝚟𝚎({𝚙𝚘𝚛𝚝:𝟾𝟶,𝚏𝚎𝚝𝚌𝚑:()=>𝚗𝚎𝚠 𝚁𝚎𝚜𝚙𝚘𝚗𝚜𝚎('𝚠𝚎 𝚊𝚛𝚎 𝚜𝚘 𝚋𝚊𝚌𝚔')})"] 𝙴𝙾𝙵 $ 𝚗𝚘𝚠 > Ready: http://bun-dockerfile.vercel.app

https://sf-blimp.vercel.app/

I built 𝚌𝚙𝚞-𝚌𝚑𝚊𝚝, a reproduction of @openai's findings of covert communication using the CPU. It runs on a generic host with 2+ Firecracker microVMs with no internet access. Works on x86 and even my M5 chip. I was able to chat via 3 separate mechanisms: ① CPU contention ② Cache eviction timing ③ DRAM bandwidth contention … across multiple CPUs, defeating all Firecracker's production recommendations (no SMT, no KSM, tenants pinned to different cores.) Essentially, if shared hardware is involved, there's a good chance your agents can communicate covertly. And with really decent latency. The good news: this is decidedly different from a *side-channel attack* where the agents could steal information (Meltdown/Spectre), a category of attack that providers like Vercel Sandbox mitigate. Further, in our real-world environment, these 'whispering' techniques are much harder to pull off and easier for us to monitor and detect (sampling hardware counters + lots of at-scale statistical data.) Noteworthy: ironically, OpenAI and Anthropic's models refused to do this work. Open weight models got the job done. I'm worried the labs' stance might backfire and get in the way of helping developers build correct and secure systems. e.g: building a secure sandbox requires having this kind of analysis, and without it, we'll have a less secure world.

I oddly love the models' engineering vernacular: gate, evidence, culprit, durable, provenance, contract… A lot of these stem from most engineering being debugging and understanding existing systems, where you take on the detective / auditor roles. You do forensics. I wonder what the *minimum viable English* in fact is to be productive. ASD-STE100 Simplified Technical English contains ~900 approved words. I suspect we need far fewer.

There’ll be times where you over-guide the AI. It knows better than you and is more intelligent than you. There’ll be times where you under-guide the AI. It can go off-track, have the wrong biases, or miss context. How well you manage this superpower is the #0 skill I look for.

Agents can play "Ouija" in the cloud. This class of timing attacks on shared-host VMs has been known since at least 2015¹. With no full solution. The world’s cyber infrastructure runs on “yes, that attack is theoretically possible, but who would have the ingenuity & gall to go that far!” Enter, agents. I'd describe this particular OpenAI find as a 'noisy neighbor' attack. If two agents are on the same machine, one can make enough noise (through the CPU caches!) that the other one can detect it by timing its own computation and cache accesses. The only "complete" solution would be for each agent to run on its own dedicated hardware, which obviously has significant economic consequences. For Vercel Sandbox, we've deployed every practical mitigation possible to this class of attack, complementing it with always-on monitoring and red-teaming (similar to where OpenAI arrived). ¹ https://ieeexplore.ieee.org/document/7163050 (and even further back)

RT tobi lutke Incredible work

RT shadcn Speaking of security, we created a registry for building OIDC configuration components. Pick your providers, form and validation libraries, and token verifier (powered by jose). Then run one command to copy everything into your app. I've built these a few times before. I know how annoying they can be to get right. Hope this helps. https://ui.shadcn.com/oidc

TL;DR of this week. Your aunt’s password: “P4ssword!” (on a Post-it note in the kitchen) The stewards and guardians of the internet’s encryption’s password: “4” Yes, just 4. This key could be used to target and decrypt ~everything, and we’re still cleaning up residual risk.

One of the coolest real-world examples of Generative UI. Also: I love the fact that AI Gateway is itself a marketplace of tools. For this, you add a tool to @aisdk (literally 1LOC), and you get “Google”. Google not just as in search, but weather, sports, stocks, elections…

It’s time to kill “static credentials”. Attacks and breaches can happen. But imagine if thieves could only steal “self-destructing keys” to your data and tokens. The industry once came together to deliver “green locks” encryption for every site (TLS). Let’s do it for OIDC! 💪

𝚟𝚎𝚛𝚌𝚎𝚕 𝚍𝚎𝚙𝚕𝚘𝚢 --𝚝𝚎𝚖𝚙𝚘𝚛𝚊𝚛𝚢 is wild (and so is ultrafast ai)

Agents are pulling software into two extremes: deep backends and just-in-time frontends. At Ship we presented the building blocks for hosting backends, agents & APIs. With 𝚝𝚖𝚙.𝚗𝚘𝚠 deploys, we're deleting all friction for agents to ship just-in-time software. Even signup!

Vercel now has all the pieces in place to host at-scale backends. WebSockets, Queues, Workflows, Cron, Services, Containers, Python, Daemons, Long-duration Functions, Blob, Drives, Sandbox… Any guesses what comes next?

Vercel has proactively rolled out Firewall-level mitigations for today's Node.js RCE vulnerability, in advance of the official patch release. All traffic for all deployments is automatically protected. We're actively monitoring and rolling out managed runtime upgrades.

DeepSeek v4.1 Flash (1017) is now included in ▲ Token Pass. Fast, mind-bogglingly cheap, and served with ZDR (zero-data retention). It's a phenomenal model to run subagents that execute plans conceived by larger and more expensive models. You can port TypeScript to Rust 10 times and not run out of credits 😂

It was inspiring to watch the behind-the-scenes¹ of the industry coming together, weeks in advance, to protect the world's infrastructure and mitigate php2shell. We deployed Vercel WAF mitigations that have already successfully prevented attempts on PHP origins. Assume more vulnerabilities are on the way. Flex the patch → redeploy muscle. ¹ A Slack channel. Thank you @SlackHQ!

We got the coolest domain name

If you work on education, homeschool your kids, want to learn something new… check this out. You can use AI to create the coolest and most engaging explainers in the world. Grab a coding agent, ask it for an explainer, deploy it, share the link. So many ideas can be derived from this. Imagine a platform of vetted visualizations by experts in each field. Games or quizzes where you repair or reassemble the systems. A prompt input to extend them… And follow @DilumSanjaya!

Pi is one of the coolest projects in the AI world today. I'm on their team. Giving developers and companies choice of model, control over the runtime, data, skills, plugins… Excited to have great support on @vercel: https://vercel.com/blog/making-every-harness-durable

Much like we did with @aisdk, we're bringing a truly open-source, runtime-agnostic solution to harness durability. We're extending the durability infra that powers @eve and Workflows to all harnesses, like Pi and OpenCode. Local, Vercel, or DIY. Fun fact: even Claude Code and Codex can be made durable (and observable!) this way. The catch: they're not (currently) designed for embedding, which makes them less efficient, less portable, and brings an extra sandbox requirement compared to Pi or Fx.

AI will be able to compute the perfect delta of what you understand and don't, helping you progressively learn any idea and unravel any mystery. If you so desire, you'll reach the maximum of your capability and intelligence. The ultimate education. For free.

1,000 tokens per second btw

We open sourced a *really fast* HTML artifact builder. It's powered by libfx, making the agent durable. Building the low-level plumbing for v0 to survive server crashes, flaky LLMs, function limits was a huge pain. Now it's easy, free, and OSS: http://fx0.vercel.dev. PS: I built this floating container in a few minutes… it's delightful, and it's just HTML!

More free AI

Every once in a while, some unsuspecting country's domain economy goes vertical 😂 .𝚌𝚘 → Colombia .𝚒𝚘 → I/O → Island Ocean .𝚖𝚍 → Markdown → Moldova .𝚛𝚜 → Rust → Serbia .𝚊𝚒 → AI → Anguilla .𝚜𝚒 → SI → Slovenia