• 

• 

Abrar AI - Professional Background Remover

A production-grade API and Dashboard for removing backgrounds from Images and Videos.

Unlike standard tools that use a single AI model, Abrar AI features an Intelligent Dual-Pipeline Engine. It automatically detects the type of image (Photo vs. Logo) and switches between Deep Learning and Precision Color Keying to ensure perfect results for every use case.

🚀 Key Features

• Auto-Detection Brain: Instantly analyzes image content to choose the best algorithm.
• AI Mode (Photos): Uses isnet-general-use (Rembg) for complex subjects like people, cars, and animals.
• Logo Mode (Graphics): Uses precision math to preserve internal details (e.g., white text inside a white logo) that AI models often destroy.
• Video Support: Full video background removal using RVM (Robust Video Matting) with smart downsampling for low-res inputs.
• Custom Backgrounds: Upload any image to automatically resize and use as the new background for your subject.
• Batch Processing: Drag & Drop 50+ images at once; the system handles the queue automatically.

---

🖼️ The "Dual-Pipeline" Difference

Standard AI models often fail on logos because they treat internal white spaces as "background." Abrar AI solves this.

1. Complex Photos (AI Mode)

Perfect for: People, Products, Animals, Real-world scenes.

• Input: A woman holding a monkey with a complex street background.
• Result: The background is cleanly removed, preserving hair and fur details.

Original	Processed

Original	Processed

2. Logos & Graphics (Logo Mode)

Perfect for: Icons, Text, Vector graphics, Solid backgrounds.

• Input: A blue "ibrar ai" logo with white text on a white background.
• Result: The white background is removed, but the white 'i' and text inside are preserved perfectly.

Original	Processed (Transparent)

---

🛠️ Installation & Local Development

Prerequisites

• Docker & Docker Compose
• (Optional) Python 3.10+ for local dev

1. Clone the Repository

git clone [https://github.com/yourusername/abrar-ai-backend.git](https://github.com/yourusername/abrar-ai-backend.git)
cd abrar-ai-backend

2. Create Environment File

Create a .env file in the root directory:

API_KEY=super-secret-key
MAX_UPLOAD_SIZE_MB=500
ALLOWED_ORIGINS=*

# Processing Settings
RMBG_IMAGE_MODEL=isnet-general-use
SOLID_BG_TOLERANCE=10
COLOR_KEY_TOLERANCE=20

# Rate Limits
RATE_LIMIT_IMAGE=1000/minute
RATE_LIMIT_VIDEO=100/minute

3. Run with Docker (Recommended)

This will start the API on port 8000 and the Dashboard on port 8090.

docker compose up -d --build

Access the App:

• Dashboard: http://localhost:8090
• API Docs: http://localhost:8000/docs

---

☁️ Deployment Guide (VPS)

Deploying to a DigitalOcean, AWS, or Azure VPS is simple.

Step 1: Prepare the Server

Install Docker and Docker Compose on your VPS.

# Ubuntu Example
apt update
apt install docker.io docker-compose -y

Step 2: Deploy Code

SSH into your VPS and clone the repo.

git clone [https://github.com/yourusername/abrar-ai-backend.git](https://github.com/yourusername/abrar-ai-backend.git)
cd abrar-ai-backend

Step 3: Configure

1. Create the .env file (copy the example above).
2. Important: Edit dashboard.html and change const API_URL to your VPS Public IP. javascript const API_URL = "http://YOUR_VPS_IP:8000";

Step 4: Launch

docker compose up -d --build

Your app is now live!

• Share this link: http://YOUR_VPS_IP:8090

---

🔌 API Usage

You can integrate this API into any app (React, Mobile, etc.).

Endpoint: POST /remove-bg

Removes background from an image.

Parameter	Type	Description
files	File	The source image (Required)
bg_file	File	A custom background image to place behind the subject (Optional)
mode	String	auto (default), ai, or color_key
background_type	String	transparent, white, black, green (Ignored if bg_file is used)

Example (Python):

import requests

url = "http://localhost:8000/remove-bg"
files = {
    'files': open('photo.jpg', 'rb'),
    'bg_file': open('beach.jpg', 'rb') # Optional
}
headers = {'X-API-Key': 'super-secret-key'}

response = requests.post(url, files=files, headers=headers)
print(response.json())

Built With

• dockerfile
• html
• python