Show HN: Frond – a frontend runtime for your app's dependency graph
14 points
3 hours ago
| 6 comments
| frondruntime.dev
| HN
killerstorm
55 minutes ago
[-]

   type ProfileSpec = Frond.NodeSpec<{
     readonly args: Frond.Args.None;
     readonly key: Frond.Key.Singleton;
     readonly deps: {
       readonly http: Frond.Dep<typeof HttpTransportNode>;
     };
    readonly result: Profile;
   }>;
This begs to be its own DSL rather than TypeScript-type-meta-programming.
reply
jaen
1 hour ago
[-]
What's the advantage of this over Jotai / atomic state / computed signals, which seems to require 10× less code with mostly the same benefits?
reply
adithyaharish
1 hour ago
[-]
Interesting approach.The lifecycle management and teardown story seems to be the main differentiator rather than state itself.How does Frond compare to Effect's Layer system? Is it essentially bringing Layer-like dependency graphs into the React runtime?
reply
alfonsodev
56 minutes ago
[-]
I understand it like events + state machine type of approach. Reacts only renders and emits events, and I guess the state machine imperatively will dictate to react what to render according to the result of executing whatever is in the state transition.
reply
chaitralikakde
51 minutes ago
[-]
How does Frond handle cyclic dependencies? Is cycle detection built into the runtime, or is it left to the developer?
reply
jatins
42 minutes ago
[-]
If I see this code in a new company I just joined I am quitting on day 1
reply
romanonthego
3 hours ago
[-]
I built Frond, a frontend runtime that makes your app's dependency graph explicit instead of leaving it scattered across provider order, enabled: user && api.ready guards, and logout teardown scripts.

The shape: your app is a graph of nodes — services, resources, screens. Each node declares its dependencies, how it's acquired, how it's cancelled, and how it's released, all in one place. The runtime resolves them in dependency order and tracks readiness, so React stays a renderer — it consumes a node that's already ready (useNode suspends until it is) instead of re-deriving that logic inside components.

Two engines run underneath. Effect handles the async work — execution guarantees, cleanup correctness, cancellation, and typed error channels. MobX handles state — granular observable state and live updates. You declare a node's dependencies, acquire, and release; Frond runs the rest on those two.

It's v0 and the API will still move.

reply
charucharu
55 minutes ago
[-]
How opinionated is Frond? Could someone adopt it incrementally, or does it need to own the entire application's dependency graph?
reply