Quick Start

The shortest path from installed package to a first model response.

1. Install

npm install aniki-sdk

2. Configure a provider

import { Aniki } from "aniki-sdk";

const apiKey = process.env.OPENAI_API_KEY;
if (!apiKey) {
  throw new Error("Set OPENAI_API_KEY before running this example.");
}

Aniki.configure({ provider: "openai", apiKey });

Aniki.configure is optional — you can also pass apiKey directly when constructing a provider — but it's the least code for a first run. If you omit apiKey entirely, the SDK falls back to the OPENAI_API_KEY environment variable.

3. Create and run an agent

import { Agent, Runner } from "aniki-sdk";

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

const runner = new Runner();
const result = await runner.run(agent, { message: "Say hello in one sentence." });

console.log(result.content);

model is passed straight through to the provider with no validation on the SDK's side, so any model identifier your provider account accepts will work here — "gpt-4o-mini" is just this guide's example.

Expected output

Hello! I'm here and happy to help with whatever you need.

The exact wording will differ every time you run it — this is the model's real, non-deterministic output, not a fixture.

Explanation

Agent only stores configuration — name, instructions, model, and which provider to use — it never talks to a provider itself. Runner.run is what actually sends the request: it loads any prior conversation history from the agent's session, calls the provider, and returns a RunResult whose content field is the model's final reply.

Continue to Your First Agent for a deeper look at what Agent and Runner are doing, or jump straight to Providers if you want to understand what's actually supported today.