Quick Start

1. The CLI (zero code)

npm run chat

you › what's the weather in Goa?
  ↪ handoff: RouterAgent → WeatherAgent
  ⚙ calling get_weather({"city":"Goa"})
  ✓ get_weather done (612ms)
WeatherAgent › It's sunny in Goa right now, around 31°C (feels like 34°C)...

Try:

• write a hello world program in c++ and run it → hands off to CoderAgent, writes hello.cpp, then asks approval before running g++ ...
• what's 12 * (4 + 7) / 2 → answered directly with the calculator tool
• /trace → toggle per-turn execution traces
• /help → list commands

2. The SDK (3 lines to a running agent)

import { Agent, Runner, OpenAIProvider, ProviderRegistry } from "archadi-agent";

ProviderRegistry.instance().register(new OpenAIProvider({ apiKey: process.env.OPENAI_API_KEY! }));

const agent = Agent.builder()
  .name("Assistant")
  .instructions("You are a helpful assistant.")
  .model({ provider: "openai", model: "gpt-4o-mini" })
  .build();

const result = await new Runner().run(agent, "Say hi in 5 words.");
console.log(result.finalOutput);

See /examples for tools, handoffs, structured output, and streaming.