Gregor

📢 I'm really excited to share that I'm back to working on Open Source, an SDK no less! And not any SDK, but @aiSDK! Thank you @vercel for letting me 🖤 I'll be using Twitter exclusively for work.

Big DX improvement for tool calling coming to @aisdk. Give Lars your feedback, now is the time to influence the API design

We've hit 10,000,000 weekly @aisdk downloads 🤯 𝚗𝚙𝚖 𝚒 𝚊𝚒 is all you need. One package, any model.

@nicoalbanese10 artisan coding! What joy

Writing code by hand today, very strange

Get started with Gemini 3.1 Flash-Lite using @aisdk! const result = streamText({ model: google('gemini-3.1-flash-lite-preview'), prompt: 'Hello World.', });

🧃 Introducing stereOS: a Linux based operating system hardened and purpose built for AI agents. It's clear that agents need an ACTUAL operating system (not what people are calling an "OS") to witness the full breadth and depth of their capabilities while mitigating the blast radius of autonomous, untrusted actors. But there are so many problems with AI sandboxes today: * Going out to the apple store and buying a mac mini will never scale and is way too expensive (obviously) * Running in Docker is too restrictive (agents can't stand up their own container infrastructure, no sub virtualization, docker-in-docker is very broken) * Firecracker strips all the hardware so GPU PCIe passthrough, secure boot, FIPs, etc. is out of the question. * Native VMs are too fat and the overhead of 1 agent per VM is too much. stereOS takes a different approach: it's a full NixOS system that you boot and then kick off agent sandboxes inside with gVisor + /nix/store namespace mounting. Each agent gets their own kernel and the /nix/store is read only by nature. Even if the agent was somehow able to escape the gVisor virtual kernel, they'd land on the NixOS system as the "agent" user! Not your actual hardware!! If you want to take a defense-in-depth approach, we support "native" agents that run at the system level kicked off by our `agentd` utility. These agents, on their own, can manage and kick off other sub agents using the internal sandboxing mechanisms. Today, we're open sourcing all of this: * stereOS: our purpose built Linux OS - github.com/papercomputeco… * masterblaster: client utility to launch, manage, and orchestrate agents - github.com/papercomputeco… * stereosd: the stereOS system control plane daemon - github.com/papercomputeco… * agentd: the stereOS system agent management daemon - github.com/papercomputeco… Give it a try, throw us a star, and let me know what you think 🧃⭐️

@Dan_The_Goodman @lgrammel merged, released, and backported to v5 as well. Please let me know if you have any trouble cc @cramforce

@Dan_The_Goodman @lgrammel Will have a look today

We gotta get this merged in @rauchg @cramforce it's the prophecy github.com/vercel/ai/pull…

Gemini Nano Banana models have been supported in @aisdk for a while, but as of recently you can use them more intuitively, via `generateImage`.

@threepointone Sure! I tried that first but run into problems so pivoted due to time constraints. Happy to give it another go, and always prefer to collab

@gr2m Will do! would you be open to a PR that uses a native Websocket instead of import('ws')? Then I could probably drop my fork altogether

use it in agents sdk today github.com/threepointone/… (forked from ai-sdk-openai-websocket-fetch, thanks @gr2m!)

@zo0r @aisdk not yet > No multiplexing support today. Use multiple connections if you need parallel runs. #connection-behavior-and-limits" target="_blank" rel="nofollow noopener">developers.openai.com/api/docs/guide…

@aisdk how can i reuse the same socket with multiple streamText?

Use OpenAI WebSockets with the AI SDK to save up to 40% time to first byte.

@SebastianB929 @aisdk we are looking into it

@aisdk Great! Will you be supporting also the realtime audio models?

And docs: #websocket-transport" target="_blank" rel="nofollow noopener">ai-sdk.dev/providers/ai-s…

here is our demo: …-sdk-openai-websocket-demo.vercel.app

That was a funn exploration and demo to build. Congrats to @OpenAI on launching the Websocket API! We are investigating a 1st class websocket support once we see more market commitment. We do see a decrease of 40%+ TTFB (Time to first Byte) for agentic interactions!

Introducing `ai-resumable-stream` for @aisdk I have stuffed everything I learned about resuming/stopping UI message streams into a small package Fully tested against `streamText` and a locally running Redis database A simple API with just three functions: start, resume, stop

@evilrabbit_ Quality.

▲

@MarcusSchiesser @openclaw @vercel @claudeai @aisdk Maybe add it to the showcase gallery? @nicoalbanese10 will know best. Thank you for sharing your work, we are all big fans!

@gr2m @openclaw @vercel @claudeai @gr2m would like to add this to @aisdk. Let me know which sub package and I'll send a PR.

After the buzz about @openclaw, I did something mean: I cut out the heart of the 🦞 and rewrote it using @vercel's AI SDK. The result: a coding agent that implements the Vercel AI Agent interface. Think @claudeai Agent SDK — without the lock-in. github.com/marcusschiesse…

@jvrsanch @rauchg We have an example, let me know if that covers what you need? page: github.com/vercel/ai/tree… api: github.com/vercel/ai/tree…

Hey @gr2m @rauchg Are stream resumption + with abort gonna be available soon (<1 month)?