Abso: The Universal TypeScript SDK for AI - One Interface, All LLMs 🐠

#llm #openai #typescript #ai

Ever wished you could switch between AI providers without rewriting your code? Meet Abso, an innovative TypeScript SDK that unifies multiple LLM providers behind a single, OpenAI-compatible interface. Whether you're using GPT-4, Claude, or any other LLM, Abso keeps your code clean, typed, and future-proof.

⭐️ If you like type-safe AI integration, consider starring our repo to support the project! ⭐️

Why Developers Love Abso

Write Once, Run Anywhere 🎯: Switch between OpenAI, Anthropic, or any other provider with a single line change. Format difference (for example with tool calls) are automatically handled for you.
100% Type-Safe 🛡️: Full TypeScript support with auto-completion for all methods, requests, and responses
Lightweight & Blazing Fast 🪶: Zero unnecessary dependencies, minimal overhead
Streaming support 🌊: Both event-based and async iteration streaming support out of the box
Extensible by Design 🔌: Easily add new providers as they emerge
Built-in Embeddings support 🔍: Ready for semantic search and advanced text analysis
Smart Token Management 📊: Accurate token counting and cost estimation included

Getting Started in 30 Seconds

First install Abso: npm install abso-ai

Then start chatting with AI:

import { abso } from "abso-ai";

const result = await abso.chat.create({
  messages: [{ role: "user", content: "What's the meaning of life?" }],
  model: "gpt-4",
});

console.log(result.choices[0].message.content);

Advanced Features

Provider Selection

Explicitly choose your preferred provider:

const result = await abso.chat.create({
  messages: [{ role: "user", content: "Hello!" }],
  model: "openai/gpt-4",
  provider: "openrouter",
});

Streaming Support

Get real-time responses:

const stream = await abso.chat.stream({
  messages: [{ role: "user", content: "Tell me a story..." }],
  model: "gpt-4",
});

for await (const chunk of stream) {
  process.stdout.write(chunk.choices[0].delta.content || "");
}