Open Source / 2025
Feather
Self-hosted image compression and conversion for teams that want control.
Feather
Design / code / deployment
Model
Self-hosted
Core
Target-size compression
Surface
UI + API
Overview
Feather is a self-hostable image optimization service built for developers and teams that need repeatable compression without third-party SaaS dependence. It covers target-size compression, format conversion, before/after comparison, and an automation-friendly REST API — all in a single Docker deployment that keeps data and workflow under your control.
The problem
Image optimization workflows are fragmented across online tools, desktop apps, and one-off scripts, making automation difficult and privacy guarantees impossible. Teams either accept SaaS lock-in or maintain brittle custom pipelines with no unified interface.
What I built
Feather unifies interactive optimization and API-driven automation in one deployment. A web UI handles interactive compression and conversion with visual feedback, while REST endpoints let CI pipelines and backend services drive the same workflow programmatically. A target-size algorithm iterates compression parameters to hit a specific file budget.
My role
Solo build spanning product design, frontend workflow, compression pipeline engineering, REST API design, and Docker deployment configuration. Every layer from UI to container entrypoint was built and shipped as a single release.
Biggest challenge
The target-size algorithm required balancing quality degradation against aggressive file budgets, with enough iteration speed to keep the interactive experience fast. Making large-file compression feel responsive without blocking the UI was the central engineering tension.
Next project
Backslash