Orchnex: Gemini-Llama Orchestration System

Overview

Orchnex is a specialized orchestration system that combines the capabilities of Google's Gemini and Meta's Llama models. This Phase 1 release focuses on optimizing the interaction between these two powerful LLMs to provide enhanced results through prompt refinement and quality control.

Orchnex in Action

System Initialization	Enhanced Prompt
Inital Response	Meta Feedback-1
Refined Response-1	Meta Feedback-2
Refined Response-2	Final Result

How to Setup

Requirements

• Google AI Python SDK

Setup

pip install google-generativeai

• Open AI

Setup

pip install openai

Setup API keys { for now we, are limited to GEMINI and LLAMA API (NIM API) }

Steps for GEMINI API

1. GEMINI API = https://aistudio.google.com/app/apikey
2. Login through your google account
3. Create an API key
4. Select your project
5. Your api key will be create, it will look like : AIzaSxxxxxxxxxxh09xxLwCA
6. Store it safe

Steps for LLAMA API using NVIDIA NIM API

1. NIM API = https://build.nvidia.com/explore/discover
2. Login or create an NVIDIA account to access
3. Here, you can also explore other various models, but for our project its llama-3_1-8b-instruct
4. llama-3_1-8b-instruct = https://build.nvidia.com/meta/llama-3_1-8b-instruct
5. Click Build with this NIM
6. Generate an API key, it will look like this : nvapi-056fxxxxxxxxxheZxxxxxxxxxxxxxxxxxxxxbg0
7. Store it safe

Clone this repo.

1. Choose a directory to clone
2. Open Terminal, and

   cd (your_directory)

3. Clone the repository

   git clone https://github.com/harshalmore31/orchnex.git

4. Open in Code Editor, eg. VS Code
5. Open the requirements.text and download the necessary requirements and dependencies
6. To test in Action Run, .src/main.py

   python src/main.py

7. Enter the saved API keys
8. Enter the prompt, go throught the whole flow, to view the interaction within two LLM

Orchnex Architecture

flowchart of orchnex of gemini and llama

Key Features

• 🤖 Dual-LLM Orchestration: Seamless coordination between Gemini and Llama
• 🔄 PromptMaster Enhancement: Automatic prompt optimization using Llama
• ✨ Phoenix Response Generation: High-quality responses via Gemini
• 📊 Quality Control Loop: Automated refinement process
• 📈 Performance Metrics: Detailed orchestration insights

Quick Start

from orchnex import MultiLLMOrchestrator, LLMConfig

# Initialize with your API keys
config = LLMConfig(
    gemini_api_key="your_gemini_key",
    nvidia_api_key="your_nvidia_key"  # For Llama access
)

# Create orchestrator
orchestrator = MultiLLMOrchestrator(config)

# Process input with visualization
result = orchestrator.process_input(
    "Explain quantum computing",
    verbose=True
)
print(result)

How It Works

1. PromptMaster (Llama)

   • Analyzes and enhances input prompts
   • Provides structured enhancement strategies

2. Phoenix (Gemini)

   • Generates high-quality initial responses
   • Refines based on feedback

3. Quality Control Loop

   • Llama analyzes response quality
   • Gemini implements refinements
   • Process continues until quality threshold met

Current Limitations

• Currently supports only Gemini and Llama models
• Requires both API keys to function
• Optimized for specific use cases

Future Development

Phase 2 will include:

• Support for additional LLM providers
• Flexible provider interface
• Custom orchestration patterns
• Advanced configuration options

Documentation

Visit our documentation for:

• Detailed setup instructions
• API reference
• Usage examples
• Performance optimization guides