Walter Korman

Minimal, powerful, and fast, choose three!

Introducing AI CLI Generate images, video, and text from your terminal. Pipe them together. Works with any agent. ai image "sunset" | ai video "animate" → Hundreds of models → Multi-model comparison → Inline previews, no native deps → AI SDK + AI Gateway

Grok 4.3 is on AI Gateway. xAI's latest model, with improved tool calling and instruction following. Try it now with 𝚖𝚘𝚍𝚎𝚕: '𝚡𝚊𝚒/𝚐𝚛𝚘𝚔-𝟺.𝟹' vercel.com/changelog/grok…

Zo Computer rebuilt their AI stack on AI SDK and AI Gateway. The results: • Latency down ~25% • 20x more reliable • New models live in production 100x faster Full breakdown ↓ vercel.com/blog/how-zo-co…

日本のみなさん、GlagStoneを見つけてくださって本当にありがとうございます。 私の1-bitゲームに反応をいただけて、とてもうれしいです。 日本語対応も進めていきます。 Thank you so much to everyone in Japan for noticing GlagStone. I’m really happy to see people reacting to my 1-bit game. I’ll keep working on Japanese localization.

@damirchelib @vercel if you get a reproduction sample let us know, please file an issue with detail at github.com/vercel/ai/issu…

Btw, GPT-5.5 through the @vercel AI gateway throws server_error 500s for certain tool definitions that work perfectly with GPT 5.4. I haven't been able to pinpoint what exactly causes this but it is driving me crazy.

These indie devs are making a game where you can play in 2D and 3D at the same time! - The world's first 5D platformer - Solve puzzles by shifting perspective - Some things only exist in one dimension It's called Screenbound. Would you play this?

“道端のタンポポ”として必死に生きるゲーム『タンポポは耐える。』が開発中。猫やカラス、ブレイクダンスおじさんなどの攻撃に耐えよう news.denfaminicogamer.jp/news/260421t 日本のさまざまな環境に適応しながら生きていく。二酸化炭素・水・光エネルギーを集め「光合成」することで生命力を高められる

Vercel Flags is now generally available. Define flags in code and manage them directly in the Vercel dashboard. vercel.com/changelog/verc…

Claude Opus 4.7 is available on Vercel AI Gateway. Optimized for long-running agents with high-res image support and an extra-high effort level. Try with: 𝚖𝚘𝚍𝚎𝚕: '𝚊𝚗𝚝𝚑𝚛𝚘𝚙𝚒𝚌/𝚌𝚕𝚊𝚞𝚍𝚎-𝚘𝚙𝚞𝚜-𝟺.𝟽' vercel.com/changelog/opus…

@SharingPsyche @HeyGarrison @SharingPsyche any more info you can offer on where (cloud provider, local, etc.) you were running, and what model/provider, so we can investigate? We aren't doing any explicit batching/chunking.

@HeyGarrison last time i checked vercel ai-gateway batches output streams of tokens so you get "chunky" output. Direct model providers did not seem to have this issue. As far as benchmarking goes it would be pretty hard to determine if it was ai-gateway or the underlying provider.

Is anyone else interested in how the ai gateways perform? Vercel, cf, open router Thinking about doing a time to first token test benchmark

Use Vercel Sandbox with the OpenAI agents SDK as an official extension. Build agents that can run code, read files, and analyze data safely inside isolated microVMs. Control the compute and data flow from your secure cloud environment.

3 months ago I started building a coding agent that runs in the cloud. It's since written every line of code I've shipped, including itself. Today, I'm open sourcing it. Introducing Open Agents.

Seedance 2.0 is now on AI Gateway. Generate photorealistic, multi-shot videos. Try it: 𝚊𝚠𝚊𝚒𝚝 𝚐𝚎𝚗𝚎𝚛𝚊𝚝𝚎𝚅𝚒𝚍𝚎𝚘({ 𝚖𝚘𝚍𝚎𝚕: '𝚋𝚢𝚝𝚎𝚍𝚊𝚗𝚌𝚎/𝚜𝚎𝚎𝚍𝚊𝚗𝚌𝚎-𝟸.𝟶', 𝚙𝚛𝚘𝚖𝚙𝚝: '𝙱𝚕𝚊𝚌𝚔 𝚝𝚛𝚒𝚊𝚗𝚐𝚕𝚎 𝚣𝚒𝚙𝚜 𝚊𝚌𝚛𝚘𝚜𝚜 𝚘𝚏𝚏𝚒𝚌𝚎' })

Today we're open sourcing open-agents.dev, a reference platform for cloud coding agents. You've heard that companies like Stripe (Minions), Ramp (Inspect), Spotify (Honk), Block (Goose), and others are building their own "AI software factories". Why? 1️⃣ On a technical level, off-the-shelf coding agents don't perform well with huge monorepos, don't have your institutional knowledge, integrations, and custom workflows. 2️⃣ On a business level, the moat of software companies will shift from 'the code they wrote', to the 'means of production' of that code. The alpha is in your factory. Open Agents deploys to our agentic infrastructure: Fluid for running the agent's brain, Workflow for its long-running durability, Sandbox for secure code execution, AI Gateway for multi-model tokens. (Because of our focus on Open SDKs and runtimes, this codebase is a gem even if you're not hosting on Vercel.) TL;DR: if you're building an internal or user-facing agentic coding platform, deploy this: vercel.com/templates/temp…

If you are building AI Agents using @aisdk, you must be familiar with this pattern of storing conversation history after each turn of the conversation and loading it back into the agent's context before each turn. This is the standard pattern for maintaining the state of agent sessions.

Vercel Sandbox is now the #1 fastest microVM-based Sandbox. We’ve also heard from customers migrating that beyond “in the lab” benchmarks, our real-world performance & reliability are second-to-none. Literally. Whether it’s 🦞 claws, coding agents, or plain compute parallelism, we think this is the foundation of the new world. Enjoy!

AI Gateway now supports team-wide Zero Data Retention (ZDR). Building safely with multiple AI models means wrestling with fragmented data policies, per-provider negotiations, and the hope that developers do not use non-complaint providers. AI Gateway changes this with team-wide ZDR. Gateway ensures your data requirements are automatically met by only routing to providers where we have negotiated ZDR agreements. Instead of managing policies provider by provider, you get one unified data policy across Claude, GPT, Gemini, and many more providers. Toggle it on in your dashboard, and all requests will route safely without touching any code: • Team-wide ZDR • Per-request controls • Disallow prompt training Move compliance to the gateway so your team can keep shipping ↓ vercel.com/blog/zdr-on-ai…

Builders are voting with traffic. This week, Novita is ranked #7 among the top AI providers on Vercel AI Gateway, alongside leaders like Anthropic, OpenAI, and Google. What makes this exciting isn’t the ranking itself. It’s what it represents: more developers are choosing fast, production-ready inference infrastructure for the shift from models to agents. We’re proud to be part of the stack powering that transition. More to build. More to scale. vercel.com/ai-gateway/lea…

As a former big-tech guy I couldn't be more proud that big-tech runs on Vercel now. Kudos to Meta for paving the way! The velocity wins of running on Vercel are no longer optional, even if, or especially when you have 20+ years of sunk-cost-fallacy infra in your mega monorepo

Try fast mode for Opus 4.6, now on AI Gateway. 2.5x faster output token speeds with the same model intelligence. Use 𝚜𝚙𝚎𝚎𝚍: '𝚏𝚊𝚜𝚝' in Anthropic provider options or set "𝚏𝚊𝚜𝚝𝙼𝚘𝚍𝚎": 𝚝𝚛𝚞𝚎 in Claude Code with AI Gateway. vercel.com/changelog/opus…