✦ Gemini AI Studio

A full-stack AI application built with Node.js and Express.js, powered by Google Gemini 2.5 Flash. Includes a built-in browser UI with four tabs — Chat, Image Analysis, Document Analysis, and Audio Transcription — backed by a production-ready multimodal REST API.

---

📑 Table of Contents

• ✨ Features
• 🚀 Quick Start
• ⚙️ Environment Variables
• 🖥️ Frontend UI
• 📡 API Reference
• 📁 Project Structure
• 🔒 Production Readiness
• 📄 License

---

✨ Features

• Multi-tab browser UI — Chat, Image, Document, and Audio tabs in a single page
• Conversational AI chat with full message history sent to Gemini on every turn
• Markdown rendering of AI responses (via marked.js)
• Drag-and-drop file upload with image preview
• Export conversation to a .txt file
• Copy-to-clipboard on every message
• Animated typing indicator while waiting for responses
• Multimodal REST API: text, image, document, and audio endpoints
• File upload handling with multer (disk storage, auto-cleanup after processing)
• Modern ESM (import/export) support
• CORS enabled, JSON-based API responses

---

🚀 Quick Start

1. Clone the Repository

git clone <your-repository-url>
cd gemini-flash-api

2. Install Dependencies

npm install

3. Configure Environment Variables

Copy .env.sample to .env and fill in your credentials:

cp .env.sample .env

GEMINI_API_KEY=your_gemini_api_key
PORT=3000

4. Start the Server

npm run start

Open your browser at http://localhost:3000 to use the UI.

---

⚙️ Environment Variables

Variable	Required	Description
GEMINI_API_KEY	Yes	Google Gemini API key
PORT	No	Server port (default: 3000)

---

🖥️ Frontend UI

The app serves a built-in UI from the public/ folder at http://localhost:3000.

Tab	Description
💬 Chat	Multi-turn conversation with Gemini. Supports markdown, copy, and export.
🖼️ Image	Upload an image (JPG/PNG/WEBP/GIF) with an optional prompt for AI analysis.
📄 Document	Upload a document (PDF/DOC/DOCX/TXT) with an optional prompt for AI analysis.
🎵 Audio	Upload an audio file (MP3/WAV/M4A/OGG) for AI transcription or analysis.

UI features:

• Drag-and-drop or click-to-browse file selection
• Image preview before submission
• Animated typing indicator (bouncing dots) while waiting
• Markdown-rendered responses with syntax-highlighted code blocks
• One-click copy on any message
• Export full chat history as a .txt file
• Fully responsive — works on mobile

---

📡 API Reference

Base URL

http://localhost:3000

---

1. Conversational Chat

POST /api/chat Content-Type: application/json

Sends the full conversation history to Gemini and returns the next AI response. Used by the Chat tab.

Body:

{
  "conversation": [
    { "role": "user",  "text": "Hello!" },
    { "role": "model", "text": "Hi! How can I help you today?" },
    { "role": "user",  "text": "Explain black holes in simple terms." }
  ]
}

Response:

{ "result": "A black hole is a region of space where gravity is so strong that nothing — not even light — can escape it..." }

---

2. Generate Text

POST /generate-text Content-Type: application/json

Body:

{ "prompt": "Explain the theory of relativity in one sentence." }

Response:

{ "result": "The theory of relativity explains how space, time, and gravity interact depending on motion and mass." }

---

3. Generate Text from Image

POST /generate-text-from-image Content-Type: multipart/form-data

Key	Type	Required	Description
image	File	Yes	Image file (.jpg, .png, etc.)
prompt	String	No	Custom instructions for the AI

Default prompt: "Please draw conclusions from the following picture."

---

4. Generate Text from Document

POST /generate-text-from-document Content-Type: multipart/form-data

Key	Type	Required	Description
document	File	Yes	Document file (.pdf, etc.)
prompt	String	No	Custom instructions for the AI

Default prompt: "Please draw conclusions from the following document."

---

5. Generate Text from Audio

POST /generate-text-from-audio Content-Type: multipart/form-data

Key	Type	Required	Description
audio	File	Yes	Audio file (.mp3, .wav, etc.)
prompt	String	No	Custom instructions for the AI

Default prompt: "Please create a transcript from the following audio."

---

✅ Standard Response

{ "result": "Generated response from the Gemini model." }

❌ Error Response

{ "error": "Error message details" }

---

📁 Project Structure

gemini-flash-api/
│
├── public/                 # Static frontend (served at /)
│   ├── index.html          # Multi-tab UI (Chat, Image, Document, Audio)
│   ├── script.js           # Frontend logic (tabs, uploads, markdown, export)
│   └── style.css           # UI styles
│
├── middleware/
│   └── upload.js           # Multer disk-storage configuration
│
├── uploads/                # Temporary uploaded files (auto-created & cleaned up)
├── index.js                # Express server & all API routes
├── .env                    # Environment variables (not committed)
├── .env.sample             # Environment variable template
├── package.json
└── README.md

---

🔒 Production Readiness

• Add request rate limiting (e.g., express-rate-limit)
• Configure CORS to allowlist specific origins
• Use PM2 or Docker for deployment
• Add request validation middleware
• Implement centralized logging (e.g., winston)
• Add authentication if exposed publicly
• Enable HTTPS behind Nginx or a reverse proxy

Example PM2 startup:

pm2 start index.js --name gemini-ai-studio

---

📄 License

Licensed under the MIT License.