Litechain

Litechain is a modern, ergonomic framework for building powerful LLM agents and AI workflows. It’s a drop-in replacement for Langchain, offering a unified API for OpenAI, Gemini, Claude, and Groq models, with seamless tool/function calling, LLM chaining, budget tracking, and more.

---

🚀 Why Litechain?

• One import, zero boilerplate: All you need is import litechain from "litechain".
• Unified API: Use OpenAI, Gemini, Claude, or Groq with the same interface.
• Simple tool integration: Register tools/functions as plain objects—no Zod schemas or extra imports.
• LLM chaining: Connect multiple LLMs for complex agent workflows with automatic routing.
• Budget & state tracking: Monitor token usage, costs, and conversation flow out of the box.
• Streaming & embeddings: Real-time streaming and custom embedding support for advanced use cases.

---

Installation

npm install litechain

Note: Also install the relevant LLM SDKs for your providers (e.g., openai, @google/genai, @anthropic-ai/sdk, groq-sdk).

---

Quickstart

import litechain from "litechain";

const llm = litechain.llm.openai({
  apiKey: process.env.OPENAI_API_KEY!,
  model: "gpt-4o-mini",
  budget: { limit: 10 }, // $10 budget
  memory: 'vector',
  embeddings: { provider: 'openai', apiKey: process.env.OPENAI_API_KEY! }
});

llm.systemprompt = "You are a helpful assistant.";

// Register a tool
llm.addTool({
  name: "add",
  description: "Add two numbers",
  parameters: {
    a: { type: "number", description: "First number" },
    b: { type: "number", description: "Second number" }
  },
  execute: async ({ a, b }) => (a + b).toString()
});

// Invoke LLM
const response = await llm.invoke("add 5 and 3");
console.log(response); // "8"

---

Documentation

• Introduction
• Getting Started
• Features
• Usage Guides
• API Reference
• Advanced Examples
• Use Cases
• Supported Providers

---

Litechain makes building AI agents simple, fast, and robust. Explore the docs to get started!