mindflare-sdk (TS/JS) 🌌

Official TypeScript/JavaScript SDK for the Mindflare AI platform. Build AI-powered chatbots with RAG (Retrieval Augmented Generation) in minutes.

🚀 Installation

npm install mindflare-sdk

⚡ Quick Start

Chat Integration (API Key Auth)

Use the Mindflare client to interact with your AI applications using RAG.

import { Mindflare } from "mindflare-sdk";

const mf = new Mindflare({
  apiKey: "mf_secret_your_api_key",
  baseUrl: "https://mindflare-api.onrender.com", 
});

// Single-shot answer
const answer = await mf.ask("How does RAG work?");
console.log(answer);

// Full conversational chat
const response = await mf.chat({
  messages: [
    { role: "user", content: "What's in my knowledge base?" },
  ],
});

Admin Management (JWT Auth)

import { MindflareAdmin } from "mindflare-sdk";

const admin = new MindflareAdmin({
  baseUrl: "https://mindflare-api.onrender.com",
});

// Authenticate
await admin.auth.login({ email: "[email protected]", password: "password123" });

// Create a knowledge base from a PDF or Website
const kb = await admin.knowledgeBases.create({
  kb_name: "Manual v1",
  source_type: "website",
  source_url: "https://docs.example.com",
});

📄 API Reference

Mindflare (Client)

• .chat(options): Multi-turn chat with history.
• .ask(question): Quick single-shot response.
• .mountChat(selector): Inject a ready-to-use chat widget into your DOM.

MindflareAdmin (Infrastructure)

• admin.auth: login(), register(), me()
• admin.apps: list(), create(), delete(), updateConfig()
• admin.knowledgeBases: list(), create(), delete(), waitForReady()

🤝 Support

Contact @srizdebnath on Discord or visit mindflareai.vercel.app.

Developed by Sriz Debnath.