Hephast
01-2025 / 09-2026
From the idea to the design, from the build to the launch. Now live on the stores. Hephast is building tomorrow's training log, with coach follow-up and nutrition gathered in one place.
Overview
Platform
Mobile / Web
Timeline
01-2025 / 09-2026
Domain
I run LightApp Limited. Hephast is its first product, and above all an idea I have been carrying for more than three years. The goal is to support people who lift with one app that brings together the training log, the coach's follow-up and the nutrition side. All of it open, so the three work in step rather than each in their own corner. Social features round out the whole.
The Solution
Free, offline, and built for every need
Hephast aims to be the hub that gathers what lifters need, from the training log to coaching and nutrition follow-up. Coach integration lands soon, along with a social space to share performances. The app is a toolbox that builds any set type you can imagine in strength training. Apple Watch support. Over 1300 exercises, and no account required. Filters and charts follow how effort progresses and show the targets left to reach.
The Challenge
Offline first drove the development. The app has to run with no account and no network, while keeping the option to sync when it is wanted. It aims to be simple, efficient and smooth, which made performance the second big subject. The third block of work covers Apple Watch support, Live Activity and driving a session from the notification area. On top of that I am alone on the project: design, marketing, development, communication, product direction. All of it on my own time, next to my job, which is no small challenge.
My work
Hephast is a project I have had in mind for a long time. I come from the military and I have always been an athlete. I lift seriously myself, and the first purpose of the app is to answer my own need. From that personal view I am trying to ship the most complete product I can, after spotting several friction points in this market. The aim is to improve the daily work of coaches, nutritionists and athletes.
The Solution
The Challenge
Training apps cost a lot, for quality that is sometimes questionable. Almost all of them force a connection and an account. What they mostly lack is coverage of the set types lifters actually use. And nobody sees anyone else's work: most coaches juggle six to eight different tools to follow a single client.
My Approach
I work in short cycles: one ticket, one branch, one pull request, reviewed before it merges. Anything I have not done before gets researched and written down as a spec first. The app, the backend and the marketing site each live in their own repository with their own release path.
The Outcome
The app reached the stores recently, after six months of intense development, alongside the marketing site. It keeps growing. Large technical blocks are still ahead, such as the coaching and nutrition integration. And I spend a lot of time improving what is already in place and fixing the bugs that show up in real use.
Key Metrics
20 months
Duration
Solo Founder
Role
iOS, Watch, Web
Platform
670
Commits
Feature
Session tracking with Focus and Detail modes, program builder, exercise library of 1300 entries, progression charts, and a companion Apple Watch app.
Offline & Sync
Local-first database on the device, Postgres backend with row level security, sync on demand through remote procedure calls. Scheduled account deletion across eight tables for GDPR erasure.
Design
Full visual identity: app icon, dark theme, typography, token system. Plus the store screenshots and the marketing visuals.
Animation
Rest timers, session intensity heatmap and performance charts drawn on the GPU. Every animated screen measured against a 55 frames per second floor.
Testing
Unit and component tests on the app, end to end coverage on the marketing site, Lighthouse budgets enforced in continuous integration.
Debug
Sentry on both the app and the backend. Commit hash embedded in every build to trace a crash back to the exact code that shipped.
CI / CD
GitHub Actions on three repositories. EAS build profiles, over the air updates, quality gates blocking merges, versioned SQL migrations on the backend.
Web & Launch
Marketing site in French and English with waitlist capture, a 1RM calculator, hreflang and structured data. Store listings, keyword research and the launch communication plan.
Ready to start your project?
Let's discuss your idea and find the best approach to bring it to life.