Christian Wu

YAP - YAP teaches you to speak a language, not tap it.

by
Language apps have a dirty secret: they teach you to tap, not talk. After 300+ days on Duolingo, most users still can't hold a conversation. YAP is built around speaking from day one. You talk, our AI listens, and gives real-time feedback on your pronunciation. Every lesson you complete is verified onchain as a Proof of Language Learning credential. Just actual speaking practice. Give us an early test as we continue to improve our product! Love, YAP

Add a comment

Replies

Best
Christian Wu
Hey everyone, Christian here, CEO and Co-Founder of YAP. I want to tell you why I built this. The real reason. When my mother was dying, I couldn't talk to her. I grew up speaking Cantonese and Spanish. When my family immigrated to the United States, I lost most of both. That's what happens. You stop using a language and it leaves you, quietly, over years, until one day you reach for it and it's not there. I took courses throughout my life. I thought I was keeping up. I wasn't. In her last moments, I had my phone out trying to look up the words I needed to say to my own mother. I couldn't tell her what she meant to me. Not in her language. Not in words that came from me. I could say the basics in Cantonese: I love you, I'm so thankful, I'll miss you. And that was it. Having your phone between you and your dying mother, searching for the words that should already be yours, is a kind of humiliation I can't really describe. After she passed, I had to process two things at once. Her death, and my own failure to speak. That was the tipping point. I ran 200+ user interviews. The same story showed up again and again, in less extreme forms. People had studied for months or years. Downloaded the apps. Hit the streaks. Then they froze in a real conversation. The problem wasn't motivation. People were motivated. The problem was that no product made them do the one thing that actually matters: open their mouth and speak. So we built YAP around one idea. You speak from the first lesson. Not after ten levels of flashcards. Not as a bonus feature. Speaking is the product. Here's the insight that most language apps get wrong: people don't need to understand every word or nail the grammar first. They need to speak. And through speaking, everything else follows. Grammar clicks when you hear yourself using it. Vocabulary sticks when you've said it out loud in context. The mouth leads. The mind catches up. Our AI listens to you in real time and breaks down your pronunciation at the phoneme level. It doesn't just say "try again." It tells you what your mouth is doing wrong and how to fix it. Then it coaches you through the correction. Less like an app, more like a patient tutor who never gets tired. We built this across 5 languages (English, Spanish, French, Italian, Portuguese) with a full CEFR-aligned curriculum. There's real progression from beginner to advanced. And every lesson you complete gets recorded onchain as a Proof of Language Learning credential. Language ability should be verifiable. Not just a line on a resume, but a real credential backed by data. Your employer, your university, or an immigration office should be able to confirm what you can actually do, not just what course you sat through. My cofounder Landon joined in the summer of 2025, after we shipped the web beta, and he's built the technical foundation that makes all of this work, from the AI pipeline to the blockchain integration on Sei. We're a small team with 500+ active users. Our 30-day retention is 43%, roughly 24x the industry average for language apps. We think that's because when people practice speaking, they feel themselves getting better. That feeling is more powerful than any streak badge. We're early. There's a lot to build. But the core thesis is holding up: make people speak from day one and they learn faster and stay longer. If you've ever felt the gap between "studying" a language and actually being able to speak it, that gap is exactly what we're here to close. Try YAP. And if someone you love speaks a language you've lost, don't wait.
Nika

This is a cool thing, I have the same problem – I cannot remember certain words or retain them, because tapping is very passive. Active using could help. P.S. Looking forward to see German language in the future :) That one has a priority for me :)

Christian Wu

@busmark_w_nika 

German! We will keep that in mind, it's a language that has been requested a few times!

Thanks for the comment and the resonating problem with current language learning! Feel free to sign up for our newsletter on goyap.ai to hear when new languages are added!

Nika

@yapchristian Yeh, they have a big economy, so possible work opportunities are motivation to learn German :D

Jan Schutte


Hi Christian, first off congratulations on the launch. I watched your video, its inspiring to see you were able to turn a difficult situation into something positive! I also like the app idea, Ive been there with duolingo. I do have two questions, 1. what languages do you support? 2. Why the addition of on-chain language tokens?

Christian Wu

@janschutte 

Hi Jan! We have 5 language tracks for now: English, French, Italian, Spanish, and Portuguese (BR).

We added tokens as a way to provide a token of your hard work (pun intended!). We wanted to make the output of your language learning into an asset, and one of the ways to do that is to make it into a digital asset!

Sergey Petrov

i tried learn spanish on duolingo and got some vocabular, but i can't speak nor i know grammar.

will try yap for sure :-)

Christian Wu

@sergeypetrov Thanks Sergey! Please try YAP! Looking forward to seeing how you like it!

Curious Kitty
On the onchain credential: what exactly is being recorded (and what is explicitly not), who is the first “verifier” you’re designing for (employers, schools, immigration, etc.), and how will you prevent it from becoming a vanity badge rather than a trusted signal?
Christian Wu

@curiouskitty 

What's recorded on-chain: wallet address, user ID, lesson ID, and performance scores.

Not recorded: personal identity, number of attempts, or anything that could be used against a learner.

Privacy by design.

Who's the first verifier we're designing for: employers hiring for roles that require demonstrable language fluency, especially in global and remote contexts. Immigration and academia have longer cycles. We're starting where the pain is most immediate and the hiring manager already has a problem we can solve today.

On the vanity badge problem: you're right that a credential is only as trusted as the assessment behind it. The reason proof of language learning (PoLL) won't become a vanity badge is because you can't fake speaking. Our evaluation layer uses AI analysis, not tap-based interactions. A learner can't click their way to a high score the way they can on Duolingo. The credential reflects actual spoken performance, which is what makes it a signal worth trusting.

We're early. But the bet is that the hardest part of building a trusted language credential isn't the blockchain, it's having an assessment rigorous enough to back it. That's what we're building first.

Karly Kingsley
Yap is great. I spent years on Duolingo and it was helpful in a lot of ways, but not in feeling truly comfortable in having a conversation.
Nourhan AbdAllah
i really love the thought behind this as someone that has had a 300+ days streak on duolingo and is yet to be able to have a conversation using that language. will definitely give it a try and congrats on ranking today!
Krystian N-Ł.

finally someone admitting the obvious

most language apps are gamified flashcard machines

if you can’t speak after 300 days, the streak was the product - not the language

making speaking the core loop just makes sense

Mykyta Semenov 🇺🇦🇳🇱

Congratulations on the launch! A year ago I suggested this idea to a couple of language-learning platforms, but they refused to implement it at the time)

Tugay Pala

Speaking-first approach is exactly what language learning needs. Most apps train you to read and tap, but real conversations are a completely different skill. How does it handle pronunciation feedback for tonal languages like Mandarin?