Skip to main content

​
Installation

core-ai is distributed as npm packages. Install the core package along with your preferred provider adapter.

​
Requirements

  • Node.js 22 or higher
  • Zod 4 (zod@^4.0.0)
  • TypeScript 5.7+ (recommended)

​
Install the core package

npm install @core-ai/core-ai

yarn add @core-ai/core-ai

pnpm add @core-ai/core-ai

​
Install a provider adapter

You need at least one provider adapter to use core-ai. Install the package for your preferred LLM provider: 
npm install @core-ai/openai

yarn add @core-ai/openai

pnpm add @core-ai/openai
The OpenAI adapter supports chat completion, streaming, structured outputs, embeddings, and image generation.

​
Add observability (optional)

core-ai provides middleware packages for tracing and monitoring model operations. See the Observability docs for full setup guides. 
npm install @core-ai/opentelemetry @opentelemetry/api

yarn add @core-ai/opentelemetry @opentelemetry/api

pnpm add @core-ai/opentelemetry @opentelemetry/api

​
Get API keys

You need an API key from your chosen provider. Set it as an environment variable: 
export OPENAI_API_KEY="your-api-key-here"

export ANTHROPIC_API_KEY="your-api-key-here"

export GOOGLE_API_KEY="your-api-key-here"

export MISTRAL_API_KEY="your-api-key-here"
Use a .env file with dotenv to manage your environment variables during development:
npm install dotenv
Then add import 'dotenv/config'; at the top of your entry file.

​
Verify installation

Create a simple test file to verify your installation:
test.ts
import { generate } from '@core-ai/core-ai';
import { createOpenAI } from '@core-ai/openai';

const openai = createOpenAI({ apiKey: process.env.OPENAI_API_KEY });
const model = openai.chatModel('gpt-5-mini');

const result = await generate({
  model,
  messages: [{ role: 'user', content: 'Hello!' }],
});

console.log(result.content);
Run it with: 
npx tsx test.ts
core-ai uses ES modules ("type": "module"). Make sure your package.json includes this or use .mts file extensions.

​
Next steps

Quickstart

Build your first chat completion in 2 minutes

Examples

Explore real-world implementation examples