Show HN: Nola – a TypeScript superset where LLM inference is a language feature

Hi HN, in a world where 100% of code is written by AI (phew, sometimes even 200%), there might still be a small percentage of us who want to write code in a more efficient way, especially when it comes to LLM interaction. I honestly tired converting my existing TS types into JsonSchemas or wrapping a function into tool object with all this bells and whistles around arguments.

And I started experimenting with an idea - how can make LLM integration feels native? How can an LLM become part of the language / compiler, instead of being just one more external API that we have to integrate through yet another SDK?

And I took the async / await / Promise idea as a starting point, ok so: 'async' - defines an asynchronous boundary, 'Promise' - represents an asynchronous intention, 'await' - resolves it.

So I started thinking, what if we could do something similar for LLMs? That led me to create Nola, a TypeScript superset built around three concepts:

'infer' - like `async`, defines the LLM inference context,

'Intent' - encapsulates the data and instructions that will be sent to the LLM,

'ask' - like `await`, resolves the `Intent` into a typed result

Right now, Intent represents two operations: extract the data and call the function (yes native TS function), but more to come soon.

Docs at: https://nola.sh/docs

Thanks for your honest feedback