RagChat SDK

A lightweight JavaScript/TypeScript SDK for integrating RAG (Retrieval-Augmented Generation) chatbots into web applications. Supports chat, PDF knowledge ingestion, and document management.

✨ Features

• 🤖 RAG-based chat interface
• 📄 Upload, list, and delete PDFs
• 🔐 API-key–based authentication
• ⚡ Works in React, Next.js, and vanilla JS
• 🛡️ Safe handling of empty backend responses

📦 Installation

npm install sakhaglobal-chatbot-rag-sdk

or

yarn add sakhaglobal-chatbot-rag-sdk

🚀 Quick Start

1️⃣ Import and initialize

import { RagChat } from "sakhaglobal-chatbot-rag-sdk";

const rag = new RagChat({
  apiBaseUrl: "http://localhost:8000",
  apiKey: "pk_live_xxx",
});

2️⃣ (Optional) Initialize LLM providers

Call this once during app startup.

await rag.init({
  geminiKey: "GEMINI_API_KEY",
  groqKey: "GROQ_API_KEY",
});

ℹ️ You can enable one or both providers depending on your backend setup.

3️⃣ Chat with your knowledge base

const response = await rag.chat("Hello!", "session_123");
console.log(response.answer);

Response format:

{
  answer: string;
  sources?: string[];
}

📄 PDF Management

Upload a PDF

const file = document.querySelector("input[type=file]").files[0];
await rag.uploadPdf(file);

List uploaded PDFs

const res = await rag.listPdfs();

console.log(res.pdfs);
/*
[
  {
    filename: "document.pdf",
    size: "1.2 MB",
    date: "2025-01-01"
  }
]
*/

Delete a PDF

await rag.deletePdf("document.pdf");

✅ Safe even if backend returns 204 No Content

🧪 Example (React)

import { useState } from "react";
import { RagChat } from "@sakhaglobal/ragchat";

const rag = new RagChat({
  apiBaseUrl: "http://localhost:8000",
  apiKey: "pk_live_xxx",
});

export default function Chat() {
  const [msg, setMsg] = useState("");
  const [reply, setReply] = useState("");

  const send = async () => {
    const res = await rag.chat(msg, "session_123");
    setReply(res.answer);
  };

  return (
    <>
      <input value={msg} onChange={(e) => setMsg(e.target.value)} />
      <button onClick={send}>Send</button>
      <p>{reply}</p>
    </>
  );
}

⚠️ Error Handling

All SDK methods throw a RagChatError on failure.

try {
  await rag.deletePdf("file.pdf");
} catch (err) {
  console.error("Operation failed:", err.message);
}