NextSoftwareGeneration
HomeAboutServicesProjectsSnippetsArticlesContact

NextSoftware Generation

Empowering businesses with cutting-edge software solutions — from mobile apps to scalable backend systems. We turn ideas into digital reality.

Quick Links

  • About
  • Services
  • FAQ
  • Contact

Legal

  • Privacy Policy
  • Terms of Service

Connect

0304 161 0238
info@nextsoftgen.com

© 2026 NextSoftware Generation. All rights reserved.

Projects

Explore my mobile apps and open-source work.

OpenMind: The Ultimate Android Benchmark Suite for OpenRouter & Free LLMs

OpenMind is a professional-grade, local-first Android application designed for developers and AI researchers to discover, interact with, and benchmark the world's most popular free AI models.

By leveraging the OpenRouter API, OpenMind transforms your mobile device into a powerful benchmarking station, providing real-time data on model efficiency, cost-to-performance ratios, and response reliability.

🚀 Why OpenMind? (The Benchmarking Edge)

In the rapidly evolving landscape of Large Language Models (LLMs), not all "free" models are created equal. As highlighted in recent industry benchmarks, factors like Time-To-First-Token (TTFT) and Tokens-Per-Second (TPS) vary wildly across providers.

AiModelsCompete eliminates the guesswork by providing:

  • Real-Time Performance Metrics: Track exactly how fast a model responds on your network.
  • Zero-Cost Verification: A strict filtering engine that validates OpenRouter's "free" tier, ensuring you never hit unexpected credits.
  • Local-First Intelligence: All your benchmarks and chat histories are stored locally using Room, allowing for deep offline analysis.

✨ Core Features & SEO Highlights

📊 Professional AI Benchmarking

  • Latency Tracking: Measure average response times in milliseconds (ms) for every interaction.
  • Throughput Metrics: Calculate Tokens Per Second (TPS) to identify the fastest models for real-time applications.
  • Consistency Analysis: Monitor response stability over multiple sessions.

🔍 Smart Model Explorer

  • Deep Filtering: Search by context length, model provider (Meta, Google, Mistral), or performance tier.
  • One-Tap Discovery: Instantly sync the latest catalog of free models like Llama 3, Gemma 2, and Phi-3.

🎮 Developer Playground

  • MVI-Driven UI: A fluid, reactive chat interface built with Jetpack Compose and Material 3.
  • History Persistence: Keep a detailed log of model outputs to compare reasoning capabilities side-by-side.

🏗️ Technical Architecture (The Developer's Choice)

Built with modern Android best practices, AiModelsCompete serves as a reference implementation for Clean Architecture and Reactive UI.

  • UI Layer: Jetpack Compose with MVI (Model-View-Intent) for predictable state management.
  • Domain Layer: Pure Kotlin use cases for business logic decoupled from the platform.
  • Data Layer:
    • Networking: Retrofit + OkHttp for high-performance API calls.
    • Persistence: Room Database for offline-first data integrity.
    • Security: DataStore for encrypted/secure local storage of API keys.
  • Dependency Injection: Dagger Hilt for a scalable and testable codebase.

🚦 Getting Started in 3 Steps

  1. Generate your API Key: Visit OpenRouter.ai to get your free-tier key.
  2. Secure Your App: Paste your key into the Settings screen within the app.
  3. Sync & Benchmark: Pull-to-refresh the Models screen and start your first benchmark session in the Playground.

📝 About the Project

OpenMind is designed for the modern AI developer who needs to stay ahead of the curve. Whether you are building an AI-powered app or just exploring the latest in open-source LLMs, OpenMind provides the metrics that matter.

Android AIOpenRouter ClientLLM BenchmarkingFree AI ModelsLlama 3 MobileGemma 2 BenchmarkJetpack Compose AIMobile LLM ExplorerAI Performance Metrics
CodeLive Demo

Chatify – Offline-First Kotlin Multiplatform Chat Application Built with Compose & MVI

Chatify is a robust Kotlin Multiplatform application showcasing a modern local-first architecture, reactive synchronization engine, and polished UI animations.
It is built using Compose Multiplatform for shared UI across Android and iOS, SQLDelight for local persistence, and the MVI (Model-View-Intent) pattern for predictable and scalable state management.

🎯 Key Features

⚡ Local-First Architecture

Chatify works fully offline by design. Every user action is persisted locally first, ensuring instant UI updates with zero latency, even without an internet connection.

🔄 Robust Sync Strategy

A powerful synchronization engine ensures data consistency between local and remote sources using state flags:

  • pendingSync → Data waiting to be synced
  • isFailed → Sync failed after retries
  • pendingDelete → Soft-deleted items awaiting server confirmation

🧠 MVI Architecture

Implements a clean unidirectional data flow using shared ViewModels across Android and iOS for predictable state handling and easier debugging.

✨ Polished UI Animations

Features a premium Typewriter Effect for assistant responses with a custom warm-up phase (1.2s) that prevents historical messages from re-animating.

🏷️ Automatic Session Naming

Intelligently generates chat titles based on the first user message in a conversation.

🔁 Reactive UI Updates

Built with SQLDelight + Coroutines Flow, ensuring the UI always reflects the latest database state in real time.

🌐 Cross-Platform Networking

Uses a shared Ktor 3 client with type-safe API requests and centralized error handling.

🏗️ Architecture Overview

Chatify follows a layered architecture designed for scalability, testability, and clean separation of concerns:

  • Presentation (MVI) → State, Events, and ViewModel orchestration
  • Domain → Business logic, UseCases, and core entities
  • Data → Repository layer, SQLDelight database, and Ktor services
  • Sync → Platform-specific background synchronization services
  • DI → Dependency Injection using Koin

🔄 Sync & Persistence Engine

Chatify ensures data reliability using a state-driven sync system:

FlagPurposeUI Behavior
pendingSyncData waiting to be syncedShows as “sending” or “pending”
isFailedSync failed after retriesDisplays retry error message
pendingDeleteSoft-deleted but not confirmedHidden from UI until confirmed

🎭 Animation Strategy: Typewriter Effect

To deliver a premium chat experience, Chatify uses a typewriter animation system with a smart state-handling mechanism:

  • Entry Phase → Waits 1.2s for database stream stabilization
  • Seeding Phase → Existing message IDs are marked as “seen”
  • Active Phase → Only new assistant messages trigger animation

This prevents historical messages from re-animating on every recomposition.

🧰 Tech Stack

  • UI Framework: Compose Multiplatform (1.10.3)
  • Language: Kotlin (2.3.20)
  • Database: SQLDelight (2.0.2)
  • Networking: Ktor (3.0.1)
  • Dependency Injection: Koin (4.2.1)
  • Serialization: Kotlinx Serialization
  • Concurrency: Kotlin Coroutines & Flow

🚀 Getting Started

🖥️ Backend

The project includes a Node.js + Express backend:

cd backend
npm install
npm start
🤖 Android
Open project in Android Studio Ladybug+
Run module: :composeApp
🍏 iOS
Open iosApp/iosApp.xcodeproj in Xcode 15+
Or run via Android Studio using Kotlin Multiplatform plugin
composekotlin multiplatform
CodeLive Demo