ImageFlow: Manage and Distribute Optimized Images

Overview

ImageFlow is a self-hosted platform for managing image collections and serving optimized versions. It handles uploads and metadata, converts images to WebP and AVIF, and can select formats and orientation based on client preferences.

It suits teams or individuals who want an image service they can run themselves, with tagging, filtering, and local or S3-compatible storage. The project combines a Go backend with a Next.js web interface and uses Redis for metadata.

Key Features

• Convert images to WebP and AVIF with configurable quality settings, while preserving GIF animation.
• Serve images with browser-based format negotiation and device-aware orientation selection.
• Upload image batches through a drag-and-drop web interface.
• Organize and filter images using tags, including exclusion filters for sensitive content.
• Store images on a local filesystem or S3-compatible storage.
• Expose public random-image and authenticated management APIs.
• Support automatic cleanup of images with configured expiry times.

Use Cases

• Run a personal, self-hosted image library with a browser interface for uploading and browsing.
• Serve randomly selected, tagged images to a website or application through the public API.
• Provide responsive image delivery for a site that benefits from format negotiation and orientation-aware selection.
• Manage images in a small team using authenticated upload, listing, tagging, and deletion endpoints.

Getting Started

git clone https://github.com/Yuri-NagaSaki/ImageFlow.git
cd ImageFlow
cp .env.example .env

Configure the required API_KEY, then start the services with docker-compose up -d. Docker and Docker Compose are required. See the README for configuration and deployment details.

Considerations

• Deployment uses multiple services: frontend, backend, and Redis. Docker Compose is the documented quick-start path.
• The README recommends at least 1 GB of RAM and sufficient storage for image files.
• S3 storage requires endpoint, region, credentials, and bucket configuration. Local storage is the default.
• The README describes a Go backend and Next.js frontend, while the repository metadata lists TypeScript as its language; the metadata language should not be taken as a complete description of the stack.