Snap2Print 🖨️📸 — Local AI Passport Photo Studio

Shahil Khan — Fri, 16 Oct 2026 06:26:19 +0000

This is a submission for the Hacktoberfest Open-Source AI Challenge Week 2: Side Hustle

What I Built

Snap2Print is an AI-powered passport photo processing tool designed to turn ordinary customer photos into print-ready passport photographs with minimal manual effort.
The application can detect multiple faces in a group photo, allowing the user to select the person they want to process. It then automatically creates an appropriate passport-photo crop while also providing manual controls for adjusting the framing.
After cropping, users can optionally remove or replace the background, apply face enhancement, and preview the final passport photo before exporting it. The application can also generate a 4×6 print-ready sheet with multiple passport photos and borders, making it suitable for physical photo-lab printing.
A key design goal was to keep the workflow fast and practical on CPU-only systems, while avoiding unnecessary AI processing whenever it is not required. This makes Snap2Print useful for local photo studios and small businesses that may not have dedicated GPUs or expensive cloud infrastructure.

Demo

The complete source code for Snap2Print is available in my public GitHub repository. The repository contains the application code, AI processing pipeline, UI, and documentation required to understand and run the project locally.

Youtube Demo Video :

Code

GitHub Repository:

shahilkhan2029 / Snap2Print

Snap2Print 🖨️📸

Local-First, Zero-API-Cost AI Passport Photo Processing & 300 DPI Print Sheet Studio

Snap2Print is an offline-capable, CPU-optimized desktop web application designed for independent photo studios, copy shops, and passport photo booths. It automates the end-to-end passport photo production pipeline—from multi-person group photos to compliant international passport crops, optional background replacement, manual retouching, face enhancement, and print-ready 4×6 inch multi-copy layouts at 300 DPI.

🌟 The Problem Snap2Print Solves

Small photo studio operators and print shops process dozens of customer passport and visa photos daily. In traditional workflows:

Recurring API Costs: Cloud background removal services (e.g., remove.bg, clipping APIs) charge $0.20 to $1.00+ per image, eroding small studio profit margins.
Customer Privacy Violations: Uploading customer and minor portraits or biometric data to third-party cloud servers raises strict legal, privacy, and data-sovereignty concerns.
Hardware Limitations: Most modern AI segmentation pipelines require expensive NVIDIA GPUs with CUDA, which…

View on GitHub

How I Built It

Snap2Print is built as a local-first AI image-processing application, designed to run on ordinary consumer hardware without requiring a cloud GPU or paid image-processing APIs.

The pipeline starts by loading and validating the customer's image, including orientation handling and image integrity checks. OpenCV YuNet is used for multi-face detection and facial landmarks, allowing the user to select a specific person from either a portrait or group photo.

After selecting a face, Snap2Print performs AI-assisted passport cropping using facial landmarks and configurable aspect-ratio presets such as 35×45 mm, 2×2 inch, and 3:4. The crop can then be manually adjusted through an interactive canvas while maintaining the required aspect ratio.

Background processing is deliberately performed only after cropping. This means the segmentation model processes the smaller portrait instead of the entire original image, reducing unnecessary CPU inference and improving responsiveness. Users can also keep the original background if removal is not required.

For background replacement, Snap2Print uses local segmentation components such as MediaPipe Selfie Segmentation and lightweight ONNX-based models, followed by mask refinement. A manual eraser is also available for correcting unwanted areas without running the AI model again.

The processed portrait can optionally go through face enhancement controls and is finally placed into a mathematically calculated 4×6 inch print layout at 300 DPI, with consistent spacing and cutting borders.

The application provides two main outputs: a single passport photo at native resolution and a print-ready 4×6 sheet containing multiple copies.

Why Does Open Innovation Matter?

Open innovation is what makes Snap2Print practical as a local-first application rather than another cloud-dependent photo-processing service.

By using open-source and open-weight AI components, I can run face detection and image segmentation directly on the user's computer. This means customer photos do not need to be uploaded to an external AI service, which is particularly valuable for sensitive identity and passport photographs.

It also removes per-image API costs and reduces dependency on external services. Instead of relying on one closed AI API for the entire workflow, Snap2Print combines different lightweight AI models with traditional image processing and manual controls.

This approach makes it possible to build a flexible pipeline where different models can be selected according to the task and available hardware. The result is a tool that can provide AI-assisted photo processing while remaining accessible to small photo studios and users with ordinary CPUs.

My Agent Session

I used an AI coding agent throughout the development of Snap2Print to iteratively design the workflow, implement features, debug issues, and refine the local AI pipeline.

No DevRelay session was saved for this submission.

My First User

My first real user was Vishal Gothwal, who runs a photo studio.

Before trying Snap2Print, he was not familiar with setting up Python environments and installing the required libraries, so I helped him configure the project and get it running locally.

Once it was set up, he tested Snap2Print using real photo-studio scenarios, including passport-photo preparation. His first reaction was essentially, “Wow.” He found the ability to select a person from a group photo, automatically crop the selected face, change the background, and generate a print-ready sheet particularly useful.

However, he also gave me an important piece of practical feedback: some photographs still require additional time for face cleaning and retouching.
This feedback helped identify an important area for further improvement. Instead of assuming that AI processing is always perfect, Snap2Print keeps manual controls such as the eraser and enhancement tools available so that a studio operator can quickly correct the output when needed.

DEV Community: Shahil Khan