Week in Review: Fluency Gym Became a Real Product
97 commits. The big one was Fluency Gym going from a rough concept to an actual product surface: schema, audio pipeline, APIs, frontend, mockups, the lot.
97 commits this week and most of the energy went into one thing: Fluency Gym.
I have had this idea in my head for a while. Repetition, speaking rhythm, sentence-level fluency practice, less passive staring at explanations. This week it stopped being an idea and started looking like a proper product.
Fluency Gym
In english-exercises.org, I shipped a ridiculous amount of the stack in one go:
- database schema with 9 new tables
- audio pipeline and infrastructure
- API endpoints
- frontend implementation
- UX mockups and demo page
- docs and technical spec so future me doesn’t have to reverse-engineer my own thinking
That was the moment the feature crossed the line from “interesting direction” to “real thing I can keep building on”.
I also wrote down the broader platform vision more explicitly. That’s been overdue. If the exercises site is turning into more of a learning system, I need the architecture and product thinking to catch up with that reality.
Firstly Academy
Over in Firstly, the work was more operational:
- My Lessons switched to accordion sections
- Scribe got Gmail-style search with a
student:operator and autocomplete - flashcards gained DeepL translations
- docs got a proper audit pass so they match the code instead of some fantasy version of it
Again, not the sexiest list. Also the sort of list that makes a tool feel grown up.
Why This Week Felt Good
There’s a nice point in product work where a feature stops feeling like a scattered pile of sub-problems and starts feeling inevitable.
That happened with Fluency Gym this week.
Not finished. Obviously not finished. But coherent now.
And once a system is coherent, shipping the next layer gets much easier:
- better practice flows
- better scoring
- better analytics
- better reasons for students to come back tomorrow
Which is the whole game really.