Documentation

• GitHub Repository: https://github.com/shuding/better-all

Note: This library is not yet indexed in DeepWiki or Context7.

better-all Library

better-all provides Promise.all with automatic dependency optimization. Instead of manually analyzing which tasks can run in parallel, tasks declare dependencies inline and execution is automatically optimized.

Installation

pnpm add better-all

Basic Usage

import { all } from "better-all";

const results = await all({
  // Independent tasks run in parallel
  fetchUser: () => fetchUser(userId),
  fetchPosts: () => fetchPosts(userId),

  // Dependent task waits automatically
  combined: async (ctx) => {
    const user = await ctx.$.fetchUser;
    const posts = await ctx.$.fetchPosts;
    return { user, posts };
  },
});

// results.fetchUser, results.fetchPosts, results.combined all typed

Key Advantage: Automatic Optimization

// Manual approach - error-prone
const [user, posts] = await Promise.all([fetchUser(), fetchPosts()]);
const profile = await buildProfile(user, posts);
const [feed, stats] = await Promise.all([
  buildFeed(profile, posts),
  buildStats(profile),
]);

// better-all - dependencies declared, execution optimized
const results = await all({
  user: () => fetchUser(),
  posts: () => fetchPosts(),
  profile: async (ctx) => buildProfile(await ctx.$.user, await ctx.$.posts),
  feed: async (ctx) => buildFeed(await ctx.$.profile, await ctx.$.posts),
  stats: async (ctx) => buildStats(await ctx.$.profile),
});

Type Inference

Results are fully typed based on task return types:

const results = await all({
  count: () => Promise.resolve(42),
  name: () => Promise.resolve("test"),
  combined: async (ctx) => ({
    count: await ctx.$.count,
    name: await ctx.$.name,
  }),
});

// TypeScript knows:
// results.count: number
// results.name: string
// results.combined: { count: number; name: string }

References

• For complex DAG patterns, see dag-patterns.md