Testing is underway
We're testing Tulpy with our first users right now and refining the experience based on what we learn.
Tulpy is now in the hands of its first real users, and honestly, it's the most useful the product has ever been to build. Nothing exposes a shaky assumption faster than watching someone actually read.
What early testing is teaching us
The biggest lesson so far: difficulty calibration is everything. When the feed sits right at a learner's i+1, they read one card after another without noticing time pass. Push it a notch too hard and they stall on translations; too easy and they get bored. We're tuning how quickly Tulpy adapts from your taps, likes, and translation requests so it finds that zone faster.
Spacing that respects real life
We're also refining the spaced-repetition schedule against real usage. People don't study in tidy daily blocks — they read on a commute, skip a day, binge on a weekend. Reviews have to survive that messiness and still resurface a word at roughly the right moment. Getting this right is less about a perfect algorithm and more about being forgiving of how humans actually behave.
What's next
We're expanding the tester group carefully, fixing the rough edges they find, and adding languages in the order the waitlist asks for them. If you want in, join the waitlist — and once you're reading, tell us where it feels off. That's the feedback that makes the next version better.