npmai 0.1.4

A lightweight Ollama(no ollama installation needed) based LangChain-compatible LLM bridge ('LLaMA-3.2','CodeLLaMA-Instruct 7B','Gemma-2-Instruct 9B','Mistral 7B Instruct','Qwen-2.5-Coder 7B','Phi-3 Medium (8B)','Falcon 7B Instruct','Baichuan-2-7B','InternLM-Chat-7B','Vicuna 7B') built by Sonu Kumar.

🚀 npmai

By Sonu Kumar Ramashish

npmai is a lightweight Python package designed to bridge the gap between users and open-source LLMs. Connect with Ollama and 10+ other powerful models instantly—no installation, no login, and no API keys required.

---

✨ Features

• 🔗 Zero Setup: No local Ollama installation or complex API signups needed.
• 🤖 Multi-Model Support: Execute prompts across 10+ open-source models simultaneously.
• 🧠 Built-in Memory: (New in v0.1.2) Native memory support—no need for external Agentic frameworks.
• 🕵️‍♂️🔍📑 RAG Frame-Work: no need to install Whisper or any model locally,no need to write code for the pdf,image,video,yt-video to text just use npmai
• ⚡ Framework Ready: Fully compatible with LangChain, CrewAI, and other orchestration tools.
• 🛠️ Universal API: Access via Python, JavaScript, C++, Java, or C.

---

🖥️ Supported Models

Model Name	Description
llama3.2	Meta's latest powerful small model
gemma-2-instruct-9b	Google's high-performance open model
qwen-2.5-coder-7b	Alibaba's elite coding assistant
mistral-7b-instruct	Versatile and efficient instructor model
phi-3-medium	Microsoft's highly capable reasoning model
And many more...	Falcon, Baichuan-2, InternLM, Vicuna

---

⚙️ Installation

Install via pip in seconds:

pip install npmai
Use code with caution.

Tip for Python 3.13+: Use py -3.13 -m pip install npmai
💡 Quick Start (Python)
python
from npmai import Ollama

# Initialize the LLM
llm = Ollama()      

# Simple invocation
response = llm.invoke("What is the future of AI?", model="llama3.2")
print(response) 

🌐 API Usage (Other Languages)
If you aren't using Python, hit our global endpoint:
POST https://npmai-api.onrender.com
🟡 JavaScript
javascript
const response = await fetch("https://npmai-api.onrender.com", {
  method: "POST",
  headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    prompt: "Hello! Who are you?",
    model: "llama3.2",
    temperature: 0.4
  })
});
const data = await response.json();
console.log(data.response);

🔵 C++
cpp
nlohmann::json payload = {
    {"prompt", "Explain quantum physics."},
    {"model", "llama3.2"},
    {"temperature", 0.4}
};
auto res = cli.Post("/llm", payload.dump(), "application/json");

🆕 Latest Update: Version 0.1.4
Now you do not need to write code for RAG tools like pdf,image,video,audio,yt-video to text and no need to load whisper and other requirements locally no local process everything on cloud in free without any signup or singin or key hurdles.
⚠️ Important Notes
Experimental Use: This project is designed for educational purposes, small-scale experimentation, and demo projects.
Scale Responsibly: For high-volume production traffic, please support the original AI researchers and infrastructure providers.
🔗 Resources
Documentation: npmai.onrender.com
API Endpoint: npmai-api.onrender.com/llm
Developed with ❤️ to make AI accessible to everyone.
Developer and Maintainer:- Sonu Kumar