OMNI: Multi-LLM AI Agent Experiment

OMNI is a sleek, single-page AI prompt interface built with React and Next.js 12. It features a dark-mode design, smooth animations, and mobile-first responsiveness. OMNI connects to a multi-agent AI orchestration server, enabling dynamic interaction across several large language models in real time.

Client Omni AI

Year 2025

Role Technical AI Product Development

Category Product Development

Product Strategy Product Research Research LLM development/Coding AI Automation CI/CD Prompting React/NextJS Design

Project Overview

OMNI is a sleek, single-page AI prompt interface built with React and Next.js. It features a dark-mode design, smooth animations, and mobile-first responsiveness. OMNI connects through a Langchain-powered multi-agent AI orchestration server via Ngrok, enabling dynamic interaction across several large language models in real time.

Multi-Agent LLM AI Integration

OMNI leverages the combined reasoning power of multiple AI models:

OpenAI GPT-4o
Google Gemini 2.5 Pro
Anthropic Claude 3.7
Mistral Codestral
Qwen 3-32B (Alibaba)

Design Process

Agents are orchestrated through Langchain, allowing sequential and role-based agent interaction for smarter, contextual responses..

Architecture & Orchestration

Langchain: Multi-agent AI orchestration logic
Ngrok: Tunnels the local Langflow server for secure external access
API Layer: POST requests handled via Postman for testing and debugging
Hosted on: Vercel (Frontend)

Features

Modern UI: Clean, dark-themed interface with high contrast and smooth feel

Responsive Design: Adapts to all device sizes

Animated Interaction: Powered by Framer Motion for smooth transitions

Tailwind CSS: Utility-first styling for consistent UI design

Tech Stack

Frontend: React, Next.js
Styling: Tailwind CSS
HTTP Client: Axios
Animations: Framer Motion
Fonts: Inter (Google Fonts)
API Tools: Postman, Ngrok
Hosting: Vercel