Not only can you pollute `then`, but you can also pollute `return`! Both behaviors are specified in the ECMAScript spec.

NextJS has an MCP and for some reason people are not aware of it. It is one of the best things they did so far. Whoever is working on it, hear me out. You are a hero 🥳

Next.js + AI devtools just ended 80% of manual UI coding I built a full AI-driven interface in <3 hours using MCP & DevTools extensions🤯 medium.com/@nakiboddin.sa… #NextJS #AI #WebDev #AICoding #React #DevTools

agentclientprotocol.com is a great abstraction for coding agents. But its design is focused on editors, not automation. coding-agent-sdk builds on ACP to work with general-purpose programs — script Claude Code, Codex CLI, Gemini CLI from any Node.js app. github.com/gaojude/coding…

next-devtools-mcp v0.3.8 fixed context bloat after screenshots and agent halting after MCP initialization

Execution ability isn't the edge anymore—AI already handles that. What actually matters is knowing what to build and why. But here's the problem: you can't practice judgment the way you practice coding. There's no clear path to get better at it. What I'm trying: documenting

Finally clear instructions!!!

I needed this for windows..

I wrote a bit about how we made v0 an effective coding agent - Dynamic system prompt - Our "LLM Suspense" framework for modifying streamed content on the fly - Our AutoFixer system for fixing various issues we've seen. A good example of how powerful the pipeline is is our

This week I stopped by Vercel’s expanded headquarters on Mission Street to mark a major investment in downtown San Francisco.   After sitting vacant for more years, this space is now home to growing teams. Vercel’s expansion reflects that San Francisco remains the epicenter of

next-devtools-mcp v0.3.10 improved Next.js doc use efficiency github.com/vercel/next-de…

Jude Gao and Vercel dropped some great research on why your skills are not being recognised the TLDR is in this video I just released youtu.be/Hy_XOB83MQc?si…

The agents-md approach for Next.js docs embedding instructs an AI agent, via a system prompt, to perform agentic RAG over the Next.js documentation for every user prompt. The docs index itself acts as a path-shortening mechanism for the agentic RAG, improving retrieval speed.

GPT 5.3 Codex has achieved the highest success rate on Next.js evals. View the leaderboard ↓ vercel.fyi/next-evals-2-5

We are also working on a way to enable this experience for existing projects as long as they are on the Canary version that supports it.