Sabitlenmiş Tweet
Francisco Moretti
1.5K posts

Francisco Moretti
@franmoretti_
Founding Engineer Alaro AI. Building https://t.co/5zxQfpkxBn (Prod-ready AI Chat template)
London Katılım Mart 2013
434 Takip Edilen1.5K Takipçiler

Today is my last day at Vercel.
I've had an absolute blast over the last year.
I got to work on a lot of my own projects that helped form the basis for our AI Cloud, helped other teams ship their great ideas, and worked alongside some of the most talented people I've ever met.
I'll have news on my next adventure for you on Tuesday but for now, wanted to reflect on my time here and the projects I worked on.
Read on if you're interested ↓

English

@franmoretti_ yes! working on this next, very possible
English

Every month I periodically see the recycled take that “something better than chat” is coming for AI. That chat is temporary.
In fact, I predict the opposite. More of our work and life will happen through chat and voice interfaces of increasingly intelligent agents. 🦞 OpenClaw is the lobster in the coal mine of this.
It’s true that as humans we consume with all the senses. When we use our internal company agent at @vercel, it can answer any question in English, but also plot data and richly visualize it. All in the chat medium. You start with any question, and receive periodic reports, much like an enterprise claw, on @slackhq. If the chat is not enough, you’re one click away from more depth and refinement on a web page.
I don’t believe web pages are going anywhere. Many will evolve to accept natural language, which is the lingua franca of AI, and stream both text and complex data. We call this Generative UI (eg: json-render.dev).
And web pages will also be crucial in enriching the conversations that happen elsewhere, whether you’re chatting with your agent on WhatsApp or Slack. We’re building chat-sdk.dev to make this universal interface as easy to deploy as possible.
English

@franmoretti_ I can't be sure, but leads to more maintainable code, which leads to fewer ridiculous terrible overflowing functions lol
English

@nicoalbanese10 Oh wow, I'm surprised by the file limit strategy resulting in less lines overall
English

@franmoretti_ Also have added line number lint rules (max 500 lines) which as helped avoid this to a certain extent
although now it does this...
x.com/nicoalbanese10…
Nico Albanese@nicoalbanese10
I asked to add a lint rule for max # of lines gpt-5.3-codex did it, then it added overrides for the offending files...lmao feel the agi 😊
English
Francisco Moretti retweetledi

Introducing just-js
Well, just-bash now has optional support for JavaScript and TypeScript execution
- Full access to the built-in virtual FS
- If enabled, access to the network via fetch
- Can execute sub-shells like CLIs
- Support for most-common fs/child_process/etc. node APIs
- Runs in QuickJS/WASM sandbox
Why use this? Just like agents love bash, they love writing JS for more complicated API processing without spamming context.
justbash.dev/?agent=Can+jus…?
English

@heiwen There's already a gateway agnostic implementation so that different gateways can be implemented chatjs.dev/docs/gateways/…
E.g. some people prefer Vercel AI SDK / Openrouter
English

@franmoretti_ It’s more like: you’re using better-auth as abstraction for auth. You use hebo-gateway as abstraction for the LLM calls.
English

@heiwen Gateways are not hosted as part of Chatjs. But you could build a custom gateway adapter for hebo
English

@franmoretti_ This is cool, like T3chat but self-hosted.
We are working on hebo-gateway, an AI gateway that you embedd into your own App. Curious whether this would be helpful for you to reduce your own boilerplate.
github.com/8monkey-ai/heb…
English

@franmoretti_ Hey this seems super interesting. How is it different from the AI sdk?
English



