BusinessAssistantHub

An enterprise AI assistant platform featuring 12 specialized business experts powered by Google Gemini API, with Lithuanian language support, real-time streaming responses, and comprehensive usage tracking.

Overview

BusinessAssistantHub is a sophisticated multi-AI assistant application that provides access to 12 domain-specific business experts, each powered by Google's Gemini 2.0 Flash Lite model. The app features real-time streaming responses, persistent chat history, token usage tracking, and a Lithuanian language interface, making it an ideal business productivity tool.

Features

12 Specialized AI Assistants

1. Business Development - Market expansion and growth strategies
2. Customer Service - Client support and satisfaction
3. E-commerce - Online sales and digital commerce
4. Data Analytics - Business intelligence and insights
5. Email Marketing - Campaign management and automation
6. Personal Growth - Professional development coaching
7. Copywriting - Content creation and messaging
8. Recruitment - Talent acquisition and HR
9. SEO - Search engine optimization strategies
10. Social Media - Digital marketing and engagement
11. Virtual Assistant - General business support
12. Sales - Revenue generation and client acquisition

Advanced Features

• Streaming Responses: Real-time typing animation as AI generates responses
• Chat History: Persistent conversation storage per assistant
• Usage Tracking: Real-time token and request counters
• Daily Quotas: 1M tokens/day, 1.5K requests/day monitoring
• Automatic Reset: Midnight quota reset
• Progress Indicators: Visual token usage bars
• Lithuanian Interface: Native language support

Technical Stack

• Framework: SwiftUI
• AI API: Google Gemini 2.0 Flash Lite (gemini-2.0-flash-lite)
• Networking: URLSession with streaming
• Reactive Programming: Combine framework
• iOS Version: iOS 15+
• Architecture: ObservableObject pattern with specialized managers
• Storage: UserDefaults for chat history and quota tracking

Key Learning Concepts

This project demonstrates:

• Streaming API integration with Server-Sent Events (SSE)
• Multi-assistant architecture with shared infrastructure
• Token usage tracking and quota management
• Persistent chat history per context
• Combine for reactive data flow
• Advanced URLSession streaming techniques
• State management across multiple views

Project Structure

BusinessAssistantHub/
├── Models/
│   ├── Assistant.swift              # Assistant definitions
│   ├── Message.swift                # Chat message model
│   └── ChatHistory.swift            # Conversation storage
├── Managers/
│   ├── ChatHistoryManager.swift     # Persistent chat storage
│   └── TokenCounterManager.swift    # Usage tracking
├── Services/
│   └── GeminiService.swift          # API integration
├── Views/
│   ├── AssistantListView.swift      # Main assistant selector
│   ├── ChatView.swift               # Conversation interface
│   └── UsageStatsView.swift         # Token/request tracking
└── BusinessAssistantHubApp.swift

API Integration

Google Gemini API

// Model: gemini-2.0-flash-lite
// Streaming: Server-Sent Events (SSE)
// Rate Limits: 1M tokens/day, 1.5K requests/day

Streaming Implementation

func streamResponse(prompt: String, assistant: String) -> AsyncStream<String> {
    AsyncStream { continuation in
        // URLSession streaming
        // Parse SSE chunks
        // Extract text deltas
        // Update UI in real-time
    }
}

Code Highlights

Chat History Manager

class ChatHistoryManager: ObservableObject {
    @Published var histories: [String: [Message]] = [:]

    func saveMessage(assistant: String, message: Message) {
        // Persist to UserDefaults
        // Update published histories
    }
}

Token Counter Manager

class TokenCounterManager: ObservableObject {
    @Published var tokensUsed: Int = 0
    @Published var requestsUsed: Int = 0
    let tokenLimit = 1_000_000  // 1M per day
    let requestLimit = 1500     // 1.5K per day

    func checkMidnightReset() {
        // Automatic daily reset logic
    }
}

Streaming UI Updates

Real-time text rendering with typing animation as tokens arrive from API.

Skills Demonstrated

• Advanced API integration with streaming
• Multi-context state management
• Persistent storage strategies
• Quota and rate limiting
• Combine reactive programming
• Async/await patterns
• Server-Sent Events parsing
• Lithuanian localization
• Professional UI/UX for chat interfaces

Usage Tracking

Daily Quotas

• Tokens: 1,000,000 per day
• Requests: 1,500 per day
• Reset: Automatic at midnight
• Display: Real-time progress bars

Token Counting

Tracks both input and output tokens for accurate quota management.

Use Cases

• Business consulting automation
• Multi-domain expert system
• Enterprise productivity tool
• AI-powered business intelligence
• Professional coaching platform
• Marketing automation assistant

Localization

• Primary Language: Lithuanian
• UI Elements: Lithuanian labels and prompts
• Assistant Responses: Configurable per assistant

Future Enhancements

• Export chat history to PDF
• Voice input/output integration
• Multi-language support expansion
• Custom assistant training
• Team collaboration features
• Analytics dashboard for assistant usage
• Cost tracking and billing integration
• Offline mode with cached responses
• Image analysis capabilities (Gemini Vision)
• Document upload and analysis

Setup Instructions

1. Obtain Google Gemini API key from Google AI Studio
2. Replace API key in GeminiService.swift
3. Build and run

API Costs

• Model: Gemini 2.0 Flash Lite
• Pricing: Free tier available
• Quotas: Check Google AI Studio for current limits

---

Author: Martynas Prascevicius Contact: mpcode@icloud.com Purpose: Advanced AI integration demonstrating enterprise assistant platform