leerob

(I work at Cursor)

We also have a CLI, if you prefer coding in the terminal. We've seen this useful for folks using JetBrains or other IDEs: https://cursor.com/cli

(I work at Cursor)

1. Checkpoints/rollbacks are still a focus for us, albeit it's less used for those working with git. Could you share the bug you saw?

2. Autocomplete for prompts was something we were skeptical of as well, but found it really useful internally to save time completing filenames of open code files, or tabbing to automatically include a recently opened file into the context. Goal here is to save you keystrokes. It doesn't use an LLM to generate the autocomplete.

3. A lot of folks don't want to juggle three AI subscriptions for coding and have found the Cursor sub where they can use GPT, Claude, Gemini, Grok models to be a nice balance. YMMV of course!

(I work at Cursor) We also support running the agent in a VM at cursor.com/agents

(I work at Cursor) The CLI supports AGENT.md as well as CLAUDE.md and Cursor Rules.

Thank you!

You don't have to use SSR: https://nextjs.org/docs/app/guides/static-exports

Thank you, I appreciate that.

Should work on Vercel, you have access to full Node.js APIs in functions.

Some examples here (public git repos): https://grep.app/search?f.lang=TSX&q=useActionState

(I'm on the Next.js team) We're working with Cloudflare on an official adapter.

https://github.com/vercel/next.js/discussions/77740

Y'all are doing amazing. Keep it up!

Not sure if you will see the reply, but please reach out lee at vercel.com and I'm happy to help.

We’re working on a bot filtering system that blocks all non-browser traffic by default. Alongside that, we’re building a directory of verified bots, and you’ll be able to opt in to allow traffic only from those trusted sources. Hopefully shipping soon.

(I work at Vercel) We moved to a transformation-based price: https://x.com/TheBuildLog/status/1892308957865111918

> Is there even a way to tell Vercel: "I only want to spend $10 a month max on this project, CUT ME OFF if I go past it."? This is crazy.

(I work at Vercel). Yes, there are soft and hard spend limits. OP was using this feature, it's called "spend management": https://vercel.com/docs/spend-management

(I work at Vercel) While it's good our spend limits worked, it clearly was not obvious how to block or challenge AI crawlers¹ from our firewall (which it seems you manually found). We'll surface this better in the UI, and also have more bot protection features coming soon. Also glad our improved image optimization pricing² would have helped. Open to other feedback as well, thanks for sharing.

¹: https://vercel.com/templates/vercel-firewall/block-ai-bots-f...

²: https://vercel.com/changelog/faster-transformations-and-redu...

If a free tier user goes over the included usage on Vercel, their site gets automatically paused. There isn't a credit card required to be put down.

Did you have issues self-hosting? https://nextjs.org/docs/app/building-your-application/deploy...

(I work at Vercel) We are primarily built on AWS hardware and our own software.

For example, a custom CDN/proxy¹, build infrastructure², compute³, WAF⁴, etc. – and then take advantage of AWS' network/fiber (like AGA⁵) and their hardware.

¹: https://vercel.com/blog/life-of-a-vercel-request-what-happen...

²: https://vercel.com/blog/a-deep-dive-into-hive-vercels-builds...

³: https://vercel.com/blog/introducing-fluid-compute

⁴: https://vercel.com/blog/preventing-infrastructure-abuse-with...

⁵: https://aws.amazon.com/global-accelerator/

Heard, we do plan to support this feature with the App Router! Right now, I would recommend using the Pages Router with Next.js if this feature is critical for you.

(I work at Vercel) Yes, it will continue to be an MIT-licensed open-source library to simplify building AI apps.

If you do a static export, you don’t have a server to optimize images. It’s just a bunch of static files. The only way to optimize the image is to point to a service through the loader.

Also, the image optimization on the default Next.js server, for example deployed to a $4 VPS, also works without Vercel.

All Next.js features work when self-hosting, including APIs and image optimization: https://nextjs.org/docs/app/building-your-application/deploy...

This is not accurate: https://nextjs.org/docs/app/building-your-application/deploy...

There still needs to be more user-friendly interfaces built on top of the raw AI materials. For many of my non-tech friends, the open text input for ChatGPT is intimidating. Yes, it can do everything, but they don't know the right questions to ask. We'll need better products that make app building more "consumer" feeling, versus understanding the intricacies of prompting.

That's awesome. Yeah, I agree with your point about people without any computer skills. We're probably still years away from that reality.

(I work at Vercel) We got rid of this and made it the same price as using Neon directly. We also added a bunch more database options (e.g. Supabase) https://vercel.com/marketplace

Thanks for writing this up. I'm on the Next.js team, so wanted to help provide more context:

> Next.js middleware run in an environment that's neither Node.js nor browser a browser environment. Some of Node’s API is supported, some isn’t.

Agreed, this is super painful and one of our top requests to fix. We're working to allow access to all Node.js APIs in Middleware. Tracking here: https://github.com/vercel/next.js/discussions/71727

> The docs mention that you should treat Server Actions like public API endpoints and Next.js 15 makes exploitation harder. However, this is still a major footgun.

Next 15 makes this much better than a standard API route. We automatically strip unused endpoints, and we add unguessable, non-deterministic IDs to secure the server actions. There's no way to get around thinking about AuthN/Z with any server APIs: https://nextjs.org/blog/next-15#enhanced-security-for-server...

> You need a 3rd party library (npm i server-only) to make sure that server code only runs on the server.

This is slightly misleading – code will run server only when you place it in a Server Component. That never makes it to the browser and you can safely fetch data there. What `server-only` does is help prevent cases where you have a shared data layer in your application, and someone might import and try to use a function designed for the server on the client. This directive then makes it error out during compile time.

Love the suggestion for trying to find ways to make this a TS hint versus waiting for compilation though. Good idea.

> Layout components can't access searchParams or pathname. We added x-search-params and x-pathname headers to incoming requests in our middleware.ts to work around this limitation.

This is intentional: https://nextjs.org/docs/app/api-reference/file-conventions/l...

What are you trying to do where you couldn't use, for example, useSearchParams in a client component for this?

> It's easy to accidentally bundle and ship server-only code to the client. There is server-only, hopefully it's going to be part of Next.js at some point.

Exactly, this is what `server-only` is for – it won't be part of Next.js core because it's designed for the broader React ecosystem, but we do recommend using it for cases like this.

> It looks like the View Transitions API is coming to React and will likely be the best solution in the future.

Indeed, View Transitions are currently already available in the latest Next.js canary and we'll be sharing more on this very soon.

Hope this all helps, thanks for writing this up!

There was some incorrect information in that post, which was later followed up on by myself (I work on Next.js) – Kent linked back in his post (https://archive.leerob.io/blog/using-nextjs)