FluidKit - write Python, get SvelteKit remote functions (feedback welcome)

[AswanthManoj]

I've always been more comfortable in Python than Node/TypeScript, but I love how Svelte handles templating and how SvelteKit thinks about data flow. When remote functions dropped, the query/form/command/prerender model mapped so naturally to Python decorators that I had to try building it.

FluidKit lets you write this:

from fluidkit import query, command, form

@query
async def get_posts() -> list[Post]:
    return await db.get_all_posts()

@command
async def like_post(post_id: int) -> None:
    await db.increment_likes(post_id)
    await get_posts().refresh()

@form
async def add_post(title: str, content: str) -> None:
    await db.insert(title, content)
    await get_posts().refresh()

It reflects on your Python functions, registers them as FastAPI endpoints, and generates .remote.ts files, so on the Svelte side it's just normal remote functions:

<script>
  import { get_posts, like_post, add_post } from '$lib/posts.remote';
</script>

{#each await get_posts() as post}
  <h2>{post.title}</h2>
  <button onclick={() => like_post(post.id)}>👍 {post.likes}</button>
{/each}

<form {...add_post}>
  <input {...add_post.fields.title.as('text')} />
  <input {...add_post.fields.content.as('text')} />
  <button>Add Post</button>
</form>

Types come from Python annotations and Pydantic models → TypeScript interfaces. Single-flight mutations, cookie forwarding, file uploads, redirects all wired up. pip install fluidkit → fluidkit init → fluidkit dev to get started, no system Node needed.

Still early, would love feedback from people who've been using remote functions, especially around patterns I might be handling wrong or things I should support differently.

Website
Docs
GitHub
PyPI