🚀 OmniDev
All-in-One AI Developer Platform

Features · Quick Start · API Reference · Architecture

OmniDev is a full-stack AI developer platform that combines DevOps automation, stealth web scraping, AI vision analysis, cloud storage management, and location intelligence into a single, cohesive application. It features a FastAPI backend with a Next.js 16 frontend using Syne, DM Sans, and JetBrains Mono.

omnidev/
├── backend/                  # FastAPI (Python 3.13)
│   ├── app/
│   │   ├── main.py           # FastAPI app + Playwright lifespan
│   │   ├── config.py         # Pydantic settings (env vars)
│   │   ├── routers/          # API endpoint definitions
│   │   │   ├── devops.py     # POST /api/devops/command
│   │   │   ├── codegen.py    # POST /api/codegen/generate
│   │   │   ├── scraper.py    # POST /api/scraper/scrape
│   │   │   ├── preview.py    # POST /api/preview/check
│   │   │   ├── vision.py     # POST /api/vision/analyze
│   │   │   ├── storage.py    # GET/POST/DELETE /api/storage/*
│   │   │   └── location.py   # GET /api/location/*
│   │   ├── services/         # Business logic
│   │   │   ├── ai_service.py       # Shared Gemini client (google-genai)
│   │   │   ├── devops_agent.py
│   │   │   ├── codegen_service.py
│   │   │   ├── context7_service.py
│   │   │   ├── scraper_service.py
│   │   │   ├── preview_service.py
│   │   │   ├── vision_service.py
│   │   │   ├── storage_service.py
│   │   │   └── location_service.py
│   │   └── schemas/          # Pydantic request/response models
│   ├── requirements.txt
│   ├── .env.example
│   └── CREDENTIALS.md        # Extra credential notes (Gemini: Google AI Studio)
├── frontend/                 # Next.js 16 (React 19)
│   ├── app/
│   │   ├── page.tsx          # Landing page (OmniDev hero, features, pricing)
│   │   ├── layout.tsx        # Root layout (Syne + DM Sans + JetBrains Mono)
│   │   ├── globals.css       # Premium dark theme + industrial/AI styling
│   │   ├── components/       # Shared components
│   │   ├── devops/page.tsx   # DevOps Agent UI
│   │   ├── codegen/page.tsx  # Code Gen UI (generate + live preview)
│   │   ├── scraper/page.tsx  # Web Scraper UI
│   │   ├── vision/page.tsx   # Vision Lab UI
│   │   ├── storage/page.tsx  # Cloud Storage UI
│   │   └── location/page.tsx # Location Services UI
│   ├── lib/api.ts            # API base URL config
│   └── package.json
├── docs/                     # Documentation
│   ├── ARCHITECTURE.md       # System design & patterns
│   ├── API.md                # Full API reference
│   ├── DESIGN.md             # Design system & Stitch screens
│   └── DEPLOYMENT.md         # Deployment guides
├── CONTRIBUTING.md           # Contribution guidelines
├── CHANGELOG.md              # Version history
├── CODE_OF_CONDUCT.md        # Community standards
├── LICENSE                   # MIT License
├── .gitignore
└── README.md                 # ← You are here

• Python 3.11+ (tested on 3.13)
• Node.js 18+ (tested with Node 22)
• npm (comes with Node.js)

git clone https://github.com/himanshu748/omnidev.git
cd omnidev

cd backend

# Create virtual environment
python -m venv .venv
source .venv/bin/activate        # macOS/Linux
# .venv\Scripts\activate         # Windows

# Install dependencies
pip install -r requirements.txt

# Install Playwright browser
playwright install chromium

# Configure environment
cp .env.example .env
# Edit .env with your credentials (Gemini key: https://aistudio.google.com/apikey)

cd frontend

# Install dependencies
npm install

Terminal 1 — Backend (port 8000):

cd backend
source .venv/bin/activate
uvicorn app.main:app --reload

Terminal 2 — Frontend (port 3000):

cd frontend
npm run dev

Open http://localhost:3000 in your browser to see the UI.
Open http://localhost:8000/docs for the interactive Swagger API docs.

Create a backend/.env file based on backend/.env.example:

# Google Gemini (required for DevOps Agent, Code Gen, Vision Lab)
# Free tier: https://aistudio.google.com/apikey
GEMINI_API_KEY=
GEMINI_MODEL=gemini-2.0-flash

# Context7 (optional — recommended for Code Gen live docs)
CONTEXT7_API_KEY=

# AWS (required for DevOps Agent + Cloud Storage)
AWS_ACCESS_KEY_ID=
AWS_SECRET_ACCESS_KEY=
AWS_DEFAULT_REGION=us-east-1

# IPInfo (optional — free tier works without a token)
IPINFO_TOKEN=

# CORS (frontend URL)
CORS_ORIGINS=http://localhost:3000

📖 Need help getting credentials? See backend/CREDENTIALS.md for AWS, IPInfo, and Context7. For Gemini, use Google AI Studio.

Web Scraper and Location Services work with zero configuration!

All endpoints are prefixed with the backend URL (default: http://localhost:8000).

GET /health
→ { "status": "ok", "service": "omnidev" }

Request Body:

{
  "message": "List my EC2 instances",
  "confirm_destructive": false
}

Response:

{
  "action": "list_ec2",
  "params": {},
  "raw_result": [...],
  "summary": "You have 3 EC2 instances...",
  "needs_confirmation": false
}

Supported Actions: list_ec2, launch_ec2, stop_ec2, terminate_ec2, list_s3_buckets, describe_security_groups

⚠️ Destructive actions (stop_ec2, terminate_ec2) require confirm_destructive: true.

Request Body:

{
  "url": "https://example.com",
  "extract": "text",
  "stealth": true,
  "wait_for": ".main-content",
  "javascript": "document.querySelector('.cookie-btn')?.click()",
  "timeout_ms": 30000
}

Extract Modes: text (inner text), html (full HTML), screenshot (full-page PNG as base64)

Response:

{
  "url": "https://example.com",
  "title": "Example Domain",
  "status_code": 200,
  "content": "...",
  "screenshot_b64": null
}

Form Data:

• image (file, required) — The image file
• mode (string) — analyze, ocr, or custom
• prompt (string) — Custom prompt when mode is custom

Response:

{
  "mode": "analyze",
  "result": "This image shows...",
  "model": "gemini-2.0-flash",
  "tokens_used": 1250
}

Request Body:

{
  "prompt": "A todo app with dark mode and local storage",
  "framework": "react"
}

Response (simplified):

{
  "files": [
    { "path": "package.json", "content": "{...}" },
    { "path": "src/main.tsx", "content": "..." }
  ],
  "instructions": "npm install && npm run dev (or run in Vercel Sandbox)"
}

The frontend Code Gen page lets you pick a framework, view generated files, copy code, download a ZIP, and open a live preview via StackBlitz.

The frontend provides a premium dark-themed dashboard for each service:

• 🏠 Landing Page — Hero section, feature cards, tech stack, process walkthrough, testimonials, pricing, FAQ
• 🤖 DevOps — Natural language command input with suggestion chips, structured result display, destructive action safeguards
• ⚡ Code Gen — Prompt + framework selector, generated file tree + code viewer, one-click copy, ZIP download, and live StackBlitz preview
• 🕷️ Scraper — URL input with demo suggestions, extraction mode toggles (text/HTML/screenshot), stealth toggle, elapsed time
• 🖼️ Vision — Drag-and-drop image upload with preview, mode pills (Analyze/OCR/Custom), token & model info
• 📦 Storage — Bucket chip selector, file browser with icons/sizes/dates, prefix filter, upload form, download links
• 📍 Location — Tabbed interface (My Location / IP Lookup / Reverse Geocode / Address Search), suggestion chips, Google Maps links

• 🌑 Premium dark theme with glassmorphism and subtle grid background
• ✨ Animated gradient orbs and micro-interactions
• 🎯 API endpoint badges showing HTTP methods (GET/POST/DELETE)
• 📱 Fully responsive on mobile, tablet, and desktop
• ⚡ JetBrains Mono for code/technical elements

You can test the backend directly using curl:

# Health check
curl http://localhost:8000/health

# Scrape a website (no auth required)
curl -X POST http://localhost:8000/api/scraper/scrape \
  -H "Content-Type: application/json" \
  -d '{"url": "https://example.com", "extract": "text"}'

# Detect location (no auth required)
curl http://localhost:8000/api/location/me

# Forward geocode
curl "http://localhost:8000/api/location/geocode?q=Eiffel+Tower&limit=3"

# DevOps command (requires Gemini + AWS keys)
curl -X POST http://localhost:8000/api/devops/command \
  -H "Content-Type: application/json" \
  -d '{"message": "List my EC2 instances"}'

Or visit http://localhost:8000/docs for the interactive Swagger UI.

This project is licensed under the MIT License — see the LICENSE file for details.

Built with ❤️ using FastAPI, Next.js, and Google Gemini