We’re building ClaviSay to help people learn English from what they already read
I’ve been thinking a lot about how disconnected language learning feels from real life.
Most English learners don’t only learn inside courses or apps. They read articles, websites, work documents, newsletters, Reddit posts, tweets, and product docs every day. But when they meet a word or sentence they don’t understand, the workflow usually breaks.
They switch to a translator.
Then maybe a dictionary.
Then maybe a note app.
Then maybe a flashcard app.
Most of the time, nothing is saved or reviewed later.
That felt like a missed opportunity.
So we’re building ClaviSay, an AI language learning assistant that turns real reading into a simple learning flow:
Import or open real English content
Understand words, phrases, and sentences in context
Save useful vocabulary and snippets
Review what you actually read later
The idea is not to replace reading with AI. It’s to make real reading easier, more understandable, and more memorable.
We started with reading because it’s where language feels most real. Vocabulary is easier to remember when it comes from something you cared enough to read, not from a random word list.
Right now, ClaviSay supports importing articles, asking AI about the content, understanding words in context, saving vocabulary, and reviewing saved words through short practice sessions.
We’re still early, and I’d love feedback from other builders and language learners:
Do you think language learning tools should focus more on real-world content instead of structured lessons?
Quick answer to your actual question first: yes, I think the real-content angle is the right bet. Vocabulary sticks better when it comes from something you chose to read and already care about, rather than a random word list. Building the whole flow around real reading is a genuine edge, not just a nicer wrapper on the same old thing.
On the site itself, the build is clean. It's fast and scores a perfect 100 on accessibility, and when I looked at the source it's tidy, semantic markup with proper alt text, not the messy output a lot of launches ship. So the foundation is solid.
My main piece of feedback is about your screenshots, because right now they're doing you a disservice. Your product is a visual thing, a side-by-side reading interface with explanations, and the screenshots are exactly where someone decides whether they get it. But they're shrunk into narrow half-width rows next to the text, so the fine UI text inside them is too small to read. I opened one at full size and it's lovely and clear: three panes, the original, the translation, and the AI assist column explaining the sentence. At the size it actually shows on the page, I can't see any of that. That's your whole pitch going invisible.
Two fixes. First, let the screenshots be bigger, give them full width and room to breathe. Second, and more important, stop showing the entire app shrunk down. Crop or zoom each image to the one thing it's demonstrating: the explanation popover, the saved word, the bilingual pair. Each screenshot should show one clear action at a readable size. Show the action, not the whole interface in miniature.
Everything else is in good shape. Make the product actually visible and the page will do a lot more selling for you.
One thing I'd be careful with:
The interesting question may not be whether people prefer learning from real-world content.
It may be what they're actually choosing when they use a tool like ClaviSay in the first place.
Those sound similar, but they can lead to very different conclusions about what should be reinforced, measured, and prioritized.
I wouldn't make that call casually this early.