Introduction

Welcome to multi-llm-ts documentation! This library provides a unified TypeScript interface for interacting with multiple Large Language Model (LLM) providers.

Why multi-llm-ts?

Building applications that leverage LLMs can be challenging when you need to support multiple providers. Each provider has its own SDK, API conventions, and capabilities. multi-llm-ts solves this by providing:

• Single API: Write once, use with any supported provider
• Type Safety: Full TypeScript support with excellent IDE experience
• Feature Parity: Consistent support for streaming, function calling, vision, and more
• Easy Migration: Switch providers without rewriting your application logic

Supported Providers

multi-llm-ts supports 10+ LLM providers including:

• OpenAI - GPT-4, GPT-3.5, and more
• Anthropic - Claude 3.5 Sonnet, Claude 3 Opus/Haiku
• Google - Gemini Pro, Gemini Flash
• Ollama - Run models locally
• Azure AI, Cerebras, DeepSeek, Groq, MistralAI, xAI, and more

See the Providers Overview for the complete list and feature matrix.

Features

Chat Completion & Streaming

Get completions from any LLM provider with support for streaming responses:

const model = igniteModel('openai', chatModel, config)
const stream = model.generate(messages)
for await (const chunk of stream) {
  console.log(chunk)
}

Function Calling

Build intelligent agents with tool use via an extensible plugin system:

model.addPlugin(new ReadFilePlugin())
model.addPlugin(new WebSearchPlugin())
const result = await model.complete(messages)

Vision Support

Analyze images with vision-capable models:

const message = new Message('user', 'What is in this image?')
message.attach({ url: 'image.jpg', mimeType: 'image/jpeg' })
const result = await model.complete([message])

Structured Output

Generate JSON with schema validation:

const schema = z.object({
  name: z.string(),
  age: z.number()
})
const result = await model.complete(messages, { schema })

Next Steps

• Install the library
• Quick Start Guide
• Learn about Providers