meselab is a young product company. We look at familiar problems from a different angle, design for the people of today and of tomorrow, and use AI where it truly multiplies what a small team can do. We are building several applications; Langify, a playful way to learn English, is our first real product candidate.
Pick the word that fits the gap.
A taste of Langify's questions. The app itself is not public yet.
Learning English is not an unsolved problem: there are courses, books and apps everywhere. But the world has changed. Attention spans are shorter, people learn on phones in small moments, and they expect learning to feel like the games and stories they already love. The way we learn has to change with them, and that is exactly what we are building for.
The same view drives everything at meselab: the first answer to a problem is rarely the best one. We look for the better one, and we design for where people are going, not only where they are today.
Classic lessons ask for long, focused hours. People today learn in a few minutes at a time, so every minute has to teach and feel rewarding.
They need many words in real context, several meanings of one word (charge a phone, charge a fee), and grammar used in real situations, not more tables.
Grammar is taught as rules to memorise. People learn faster from seeing a situation and choosing what fits it.
Every scene, example and exercise is hand-made, so small teams cannot keep a course fresh, deep and correct. AI changes that, if it is used carefully.
We combine game design with an AI content pipeline, so a tiny team can ship a course that keeps getting deeper.
Tiny pixel stories with no words. Pip asks Bo for his pen, then gives it back. You name what happened: borrow. Meaning before translation.
A scene plays and stops at a NOW marker. Before it, during it, or after it? You pick, fix or build the sentence that fits. No rule tables, a hint only after a mistake.
One word, many meanings, each with its own progress. Words come from your interests, with a share from outside them, so you also discover new things.
Every word, scene and grammar case is a plain file built from a shared library of pixel art and actions. That is what lets AI agents create new lessons safely, with no designer in the loop.
A language course needs thousands of correct, varied examples. Writing them by hand is what keeps learning apps shallow. We use Claude as the author and builder, and a deterministic referee as the judge, so we get the scale of AI without letting a wrong answer reach a learner.
A Claude agent follows a project skill to write words, meanings, scenes or grammar cases, using only the existing art and actions.
An automatic checker rejects anything wrong: unknown art, duplicates, unclear answers, captions beyond basic words, and says which line to fix.
Every addition is reviewed like code before it ships. Agents propose; people decide.
The game asks it in every question type straight away, and play data shows which content works and what to write next.
Langify is developed with Claude Code: features, tests, docs and tools. A very small team ships what would normally need a large one, with an automated test suite guarding every change.
Agents extend the course on their own from written guides, a visual asset catalog and an inventory of what already exists, so they never repeat it.
Grading. The server decides every answer from checked content, and answers never reach the browser. A learner is never marked wrong by a guess.
Where Langify stands in October 2026, before launch. Every scene, case and asset is a file agents can learn from and add to.
Our first real product candidate, and the first proof of how meselab works.
A web app for learners at intermediate level and above. A placement test sets your starting point; after that, a short Today session each day. The features below are built and being tested before the first public release.
Every word, and every meaning of a word, climbs four stages. A right answer flashes its XP and moves on. A wrong one stays to teach.
Reviews that are due, a few new words picked from your interests, and grammar to practise. Casual, Regular or Hardcore pace.
Synonym, sentence meaning, fill the gap, odd one out, true or false, Type It, which meaning, and scenes.
Wordless pixel stories that teach one word each.
A grammar map: Time Machine, Where Is It?, Can You?, with stars and boss stages.
Unlimited ten-question runs of any type, or a mix.
Challenge a friend live. Same questions, every right answer hits for 10 HP.
A live knockout for 3 to 16 friends, with a champion.
Add friends by name or link, compare stats, keep streaks, earn XP and levels.
Spend coins on hats, accessories and duel effects for your pixel character.
Missed a day? Repaint it with an extra round. Finish the picture and add it to your collection.
0 / 144 DAYS
I am a young founder, and I do not want to get stuck on classical approaches, not even on just adapting to today. A good product's audience is both today's people and tomorrow's, so we have to predict how people will think, learn and work, not only follow how they do now.
That is why I started meselab: to look at problems from a different angle and build the better solution, using AI where it makes a real difference. Langify is the first step.
Mehmet Gönenç, founder of meselab
meselab
The name joins mesela, Turkish for “for example”, with lab. Every idea we have becomes a working example you can open, play and share.
We are not here to make just a language app. meselab works on different applications, each starting from a familiar problem and a question: is there a better way to solve this for people today and tomorrow? We answer it by building a complete product instead of a slide deck. Langify is the first one we are taking to launch.
We ship a working example before we write a plan, then let people tell us what is wrong with it.
AI creates at scale; checks and people decide what is true. Learners only ever see checked content.
Built on open data and open tools, with content in plain files people and agents can both add to. Made for friends first, where feedback is honest.
Want to try it first, or talk about what we are building? Write to us at mehmetgonenc@meselab.com and we will invite you when early access opens.