JSPM – @funkai/agents@0.2.1

Package Exports

@funkai/agents

Readme

@funkai/agents

Lightweight workflow and agent orchestration framework built on the Vercel AI SDK.

Features

⚡ Functions all the way down — agent, tool, workflow are functions that return plain objects.
🧩 Composition over configuration — Combine small pieces instead of configuring large ones.
🛡️ Result, never ~~panic~~ throw — Every public method returns Result<T>. Pattern-match on ok instead of try/catch.
🔒 Closures are state — Workflow state is just let variables in your handler.
🔍 $ is optional sugar — The $ helpers register data flow for observability; plain imperative code works too.

Install

npm install @funkai/agents

Usage

Agent

import { agent } from "@funkai/agents";

const helper = agent({
  name: "helper",
  model: "openai/gpt-4.1",
  system: "You are a helpful assistant.",
});

const result = await helper.generate("What is TypeScript?");

if (!result.ok) {
  console.error(result.error.code, result.error.message);
} else {
  console.log(result.output);
}

Tool

import { tool } from "@funkai/agents";
import { z } from "zod";

const fetchPage = tool({
  description: "Fetch the contents of a web page by URL",
  inputSchema: z.object({ url: z.url() }),
  execute: async ({ url }) => {
    const res = await fetch(url);
    return { url, status: res.status, body: await res.text() };
  },
});

Workflow

import { workflow } from "@funkai/agents";
import { z } from "zod";

const research = workflow(
  {
    name: "research",
    input: z.object({ topic: z.string() }),
    output: z.object({ summary: z.string(), sources: z.array(z.string()) }),
  },
  async ({ input, $ }) => {
    let sources: string[] = [];

    const data = await $.step({
      id: "fetch-sources",
      execute: async () => findSources(input.topic),
    });
    if (data.ok) sources = data.value;

    const analysis = await $.agent({
      id: "summarize",
      agent: summarizer,
      input: { text: sources.join("\n") },
    });

    return {
      summary: analysis.ok ? analysis.output : "Failed to summarize",
      sources,
    };
  },
);

const result = await research.generate({ topic: "Effect systems" });

Streaming

const result = await helper.stream("Explain closures");

if (result.ok) {
  for await (const chunk of result.stream) {
    process.stdout.write(chunk);
  }
}

API

Export	Description
`agent(config)`	Create an agent. Returns `{ generate, stream, fn }`.
`tool(config)`	Create a tool for function calling.
`workflow(config, handler)`	Create a workflow with typed I/O and tracked steps.
`createWorkflowEngine(config)`	Create a workflow factory with shared configuration and custom steps.
`openrouter(modelId)`	Shorthand to create an OpenRouter language model from env key.
`createOpenRouter(options?)`	Create a reusable OpenRouter provider instance.

Documentation

For comprehensive documentation, see docs/overview.md.

License

MIT