Examples Overview

This page provides an overview of all available examples in the core-ai repository. Each example is a complete, runnable TypeScript file that demonstrates specific features and use cases.

Running Examples

All examples are located in the examples/ directory of the repository.

Prerequisites

Node.js 22+
Dependencies installed from the repository root:

npm install

Environment Variables

Create a .env file at the repository root:

OPENAI_API_KEY=your_openai_api_key
ANTHROPIC_API_KEY=your_anthropic_api_key
GOOGLE_API_KEY=your_google_api_key
MISTRAL_API_KEY=your_mistral_api_key
OMNIFACT_API_KEY=your_omnifact_org_api_key
KIMI_API_KEY=your_kimi_api_key
AZURE_OPENAI_API_KEY=your_azure_openai_api_key
AZURE_OPENAI_ENDPOINT=https://your-resource.openai.azure.com
GOOGLE_VERTEX_PROJECT=your_google_cloud_project_id

You only need API keys for the providers you plan to use. Each example specifies which key it requires.

Run an Example

From the repository root:

npx tsx examples/01-chat-completion.ts

Basic Features

Chat Completion

Basic generate() chat completion with OpenAI. Demonstrates the simplest way to get started with core-ai.

Streaming

Streaming output with stream() and ChatStream. Shows how to handle real-time text generation with event-based streaming.

Multi-Modal Input

Multi-modal input using text and image URLs. Demonstrates how to send images alongside text for vision-enabled models.

Error Handling

Handling CoreAIError and ProviderError. Shows proper error handling patterns for both local validation and provider errors.

Advanced Features

Tool Calling

Tool definition and full tool-call round trip. Complete example of defining tools with Zod schemas and handling tool execution.

Embeddings

Embeddings with embed(). Shows how to generate vector embeddings for text using OpenAI’s embedding models.

Image Generation

Image generation with generateImage(). Demonstrates creating images from text prompts with configurable size options.

Structured Outputs

Typed structured output with generateObject(). Shows how to generate strongly-typed JSON objects validated with Zod schemas.

Streaming Objects

Streaming structured output with streamObject(). Demonstrates real-time streaming of structured JSON with incremental validation.

Stream Abort

Abort an in-progress stream with AbortSignal and handle StreamAbortedError.

Provider Examples

core-ai supports multiple AI providers with the same unified API.

Anthropic Provider

Using Anthropic (Claude) with the same generate() API. Shows how to switch providers while keeping the same code structure.

Google GenAI Provider

Using Google GenAI (Gemini) with the same generate() API. Demonstrates provider portability with Google’s models.

Mistral Provider

Using Mistral with the same generate() API. Shows how easily you can integrate Mistral’s models into your application.

Omnifact Provider

Using the Omnifact API Gateway with the same generate() API. Connect with your organization API key and any enabled model.

Azure OpenAI Provider

Using Azure OpenAI deployments with the same generate() API. Pass your deployment name as the model id.

Vertex AI Anthropic Provider

Using Claude on Vertex AI with the same generate() API. Authenticates via Application Default Credentials.

Google on Vertex AI

Using Gemini on Vertex AI with the same generate() API. Authenticates via Application Default Credentials.

Kimi Provider

Using the Kimi API with reasoning output and the same generate() API.

Example Categories

1. Getting Started

Start with these examples if you’re new to core-ai:

01-chat-completion.ts - Your first chat completion
02-streaming.ts - Real-time streaming responses
07-error-handling.ts - Proper error handling patterns

2. Advanced Capabilities

Explore powerful features:

03-tool-calling.ts - Enable models to call functions
05-embeddings.ts - Generate vector embeddings
06-image-generation.ts - Create images from text
04-multi-modal.ts - Process text and images together

3. Structured Data

Work with typed, validated outputs:

16-generate-object.ts - Generate structured JSON
17-stream-object.ts - Stream structured data

4. Stream control

Manage in-flight streams:

18-stream-abort.ts - Abort a stream with AbortSignal

5. Multi-Provider

Experience true provider portability:

08-anthropic-provider.ts - Claude models
09-google-genai-provider.ts - Gemini models
10-mistral-provider.ts - Mistral models
11-omnifact-provider.ts - Omnifact API Gateway models
12-azure-openai-provider.ts - Azure OpenAI deployments
13-anthropic-vertex-provider.ts - Claude on Vertex AI
14-google-vertex-provider.ts - Gemini on Vertex AI
15-kimi-provider.ts - Kimi models

Next Steps

Guides

Read detailed guides for each feature

API Reference

Explore the complete API documentation

GitHub Repository

Browse the source code and examples

Quickstart

Get started with core-ai in minutes

​Running Examples

​Prerequisites

​Environment Variables

​Run an Example

​Basic Features

Chat Completion

Streaming

Multi-Modal Input

Error Handling

​Advanced Features

Tool Calling

Embeddings

Image Generation

Structured Outputs

Streaming Objects

Stream Abort

​Provider Examples

Anthropic Provider

Google GenAI Provider

Mistral Provider

Omnifact Provider

Azure OpenAI Provider

Vertex AI Anthropic Provider

Google on Vertex AI

Kimi Provider

​Example Categories

​1. Getting Started

​2. Advanced Capabilities

​3. Structured Data

​4. Stream control

​5. Multi-Provider

​Next Steps

Guides

API Reference

GitHub Repository

Quickstart

Running Examples

Prerequisites

Environment Variables

Run an Example

Basic Features

Advanced Features

Provider Examples

Example Categories

1. Getting Started

2. Advanced Capabilities

3. Structured Data

4. Stream control

5. Multi-Provider

Next Steps