Polyglot SDK

Polyglot is a powerful and flexible TypeScript SDK for interacting with multiple Language Model APIs. It provides a unified interface for various LLMs, along with advanced features like streaming responses, caching, middleware support, and error handling.

Features

• Multi-model Support: Easily integrate and switch between different LLM providers (supports Claude, ChatGPT, Mistral, and Gemini).
• Streaming Responses: Get real-time responses from supported models.
• Caching Layer: Improve performance with built-in caching for repeated queries.
• Middleware System: Customize request/response processing with a flexible middleware system.
• Robust Error Handling: Automatically retry on transient errors and provide clear error messages.
• TypeScript Support: Full TypeScript support for improved developer experience.
• Extensible Architecture: Easily add new models or extend functionality.
• Rate Limiting: Built-in rate limiting to prevent API quota exhaustion.

Installation

npm install polyglot-sdk

Quick Start

import { Polyglot, Models, ModelConfig } from "polyglot-sdk";

const polyglot = new Polyglot();

const claudeConfig: ModelConfig = {
  apiKey: "your-claude-api-key",
  model: "claude-3-5-sonnet-20240620",
};
polyglot.addModel("claude", new Models.ClaudeModel(claudeConfig));

async function getResponse() {
  const messages = [{ role: "user", content: "Hello, how are you?" }];
  const response = await polyglot.generateResponse("claude", messages);
  console.log(response.content);
}

getResponse();

Supported Platforms and Models

| Platform | Supported Models | | ------------------ | ------------------------------------------------------------------------------------------------------ | | Claude (Anthropic) | claude-3-5-sonnet-20240620, claude-3-opus-20240229, claude-3-sonnet-20240229, claude-3-haiku-20240307 | | ChatGPT (OpenAI) | gpt-4o, gpt-4o-mini, gpt-4-turbo, gpt-3.5-turbo | | Mistral AI | open-mistral-nemo, mistral-small-latest, mistral-medium-latest, mistral-large-latest, codestral-latest | | Google Gemini | gemini-1.5-pro, gemini-1.5-flash, gemini-1.0-pro |

Advanced Usage

Streaming Responses

const stream = polyglot.generateStreamingResponse("claude", messages);
stream.on("data", (chunk) => console.log("Received chunk:", chunk.toString()));
stream.on("end", () => console.log("Stream ended"));

Using Middleware

polyglot.use(async (messages, next) => {
  console.log("Sending messages:", messages);
  const response = await next(messages);
  console.log("Received response:", response);
  return response;
});

Caching

Caching is enabled by default. To disable:

const response = await polyglot.generateResponse("claude", messages, {
  useCache: false,
});

Switching Models

You can easily switch between different models of the same platform:

const chatgptModel = polyglot.getModel("chatgpt");
chatgptModel.setModel("gpt-4-turbo");

API Reference

For detailed API reference, please refer to the TypeScript definitions in the types.ts file.

License

This project is licensed under the MIT License.