CapWords
Learn a new language from familiar things.
157 followers
Learn a new language from familiar things.
157 followers
CapWords helps you learn new languages using the world around you. Snap a photo of any object, and get instant translations, pronunciation, and examples. No flashcards—just real-life learning from familiar things.
This is the 2nd launch from CapWords. View more
CapWords Speak
Launching today
CapWords Speak brings language practice into your real life. Turn today's photos into instant conversations. Chat with context-aware AI buddies who remember your vocabulary, adapt to your level, and conquer speaking anxiety with zero pressure.
Free Options
Launch Team
CapWords
Hey Product Hunt! I'm Acelee, the creator of CapWords.
Today, our new feature, CapWords Speak, is finally live.
As I sat down to ship this update, I wanted to take a moment to write down the honest story and the philosophy behind why we spent 7 months building it.
THE QUESTION
To me, "Speak" was never just about plugging into a voice API and wrapping an AI tutor.
The core question we wanted to answer was simple:
How can we naturally learn a new language from the real world?
WHY IT TOOK 7 MONTHS
In the past year, we designed and built a lot of features for CapWords.
Some were close to being finished, but at the last minute, I decided to kill them.
Not because they weren't cool, but because I kept asking myself one question: Does this actually help users learn and speak a language better?
If the answer wasn't a resounding yes, I couldn't bring myself to launch it.
Last December, we were fully prepared to build a feature that auto-generated a daily story based on your day. The specs were done.
But right before we started coding, I realized it was mostly a novelty.
It wouldn't solve the exact moment when language learners freeze up and just can't open their mouths.
So, out of nowhere, I told my engineer: "We shouldn't do this. Let's build Speak instead."
Once I believe something is worth doing, I take action immediately, because action brings answers.
That choice started a grueling 7-month development journey.
THE PROBLEM
Most language apps do a great job with standard, simulated scenarios:
Ordering food at a fake cafe
Checking into a hotel
Airport dialogue scripts
Saying things like, "Chicago is very windy."
But I’ve never even been to Chicago.
When I stood in front of those practice prompts, I froze.
It wasn't because I knew absolutely nothing; it was because I had no real reason to say those sentences. They had nothing to do with me.
I wanted my practice to start from things I actually experienced: the sound of the rain outside my window right now, a street sign I saw on my commute, or those tiny, specific moments I actually wanted to describe.
THE AI GAP
AI knows the entire internet. But it doesn't automatically know you.
It doesn't know what photos you took today. It doesn't know what made you anxious or excited.
It doesn't know your physical sensations, emotions, or state of mind in the real world—the very things that are incredibly valuable to a human being.
Language belongs to a specific person's real, lived reality.
It happens when you look at a cup of coffee, walk into a local shop, try to share a small moment, or express your mood. Language is not just a tool; it's the ability to observe and understand the world.
WHAT WE BUILT
We didn't want to build just another AI that can talk—that part is easy now.
We wanted an AI that deeply understands your reality, so you actually want to open your mouth.
CapWords Speak is built around four core elements:
Continuous Memory System: Speak carries valuable context from your past chats into future ones. It distills personal details, recurring themes, and settings from your life, so it stops feeling like a one-off prompt and feels more like a continuous relationship.
Vibe-Matching Buddies: Language practice shouldn't feel like an exam. You can choose a gentle, encouraging companion, someone who talks straight to the point, or whatever style lowers your anxiety and makes speaking feel like a normal conversation.
Magic Album Context: With your explicit permission, CapWords recognizes your photos from today with one tap, turns them into word stickers, understands your day, and kicks off a natural conversation.
A True Learning Loop: See a word during a chat -> turn it into a word sticker -> use it in conversation -> save new phrases -> study them later. This is a natural cycle where points connect into lines, and lines weave into a network.
The Bigger Picture
We believe the future of AI education is highly contextual. Cameras, photos, locations, and wearables will help AI better understand our physical world.
But this must be built on explicit user consent, user control, and absolute transparency.
AI shouldn't spy. It should only understand when you allow it to. When AI understands the real world, language learning will change.
It will start from what you actually saw, what you went through, and the exact moment you felt the urge to express yourself.
Finally
Speak is not yet in its ultimate, perfect form. We refactored it completely several times, optimizing prompts and UX.
What we are launching today is simply our first step toward a deep belief:
Language learning shouldn't revolve around grammar points. It should revolve around your life.
If you’ve ever felt awkward trying to speak a new language because you just didn't know where to start, I highly recommend giving CapWords Speak a try.
And if you like it, please share it with your friends who are learning languages too.
I’ll be around all day. Please share your thoughts, feedback, or your own language learning struggles in the comments! 🙏
➡️ Try it here: capwords.app