All Products
Active

What if you stopped learning words, and started learning sentences?

I built OpenSen to explore a different way of learning a language: real situations, whole sentences, and reusable chunks instead of isolated words.

  • AI
  • Language Learning
  • React
  • PWA
  • Spaced Repetition
  • Personal Project

/01. The problem

The way we learn language is backwards.

Most learners know thousands of words but still struggle to produce a simple sentence in real time.

Traditional learning

  1. word
  2. grammar
  3. sentence
  4. conversation

Starts from isolated words and rules, so they are hard to use in a real situation.

OpenSen

  1. situation
  2. sentence
  3. chunk
  4. automatic recall

Starts from a real situation and trains reusable chunks for natural speech.

/02. Core idea

Learn language as chunks.

People remember and use language in chunks, not isolated words. OpenSen turns each sentence into a reusable frame with a slot you can swap.

A real sentence → smaller chunks you can actually reuse.

Could you help mePolite request
findAction
the station?Slot · place

Keep the frame, swap the slot

Could you help me find [PLACE]?

  • the station
  • the check-in counter
  • gate 12
  • baggage claim

/03. The system

A complete loop for language learning.

From a real situation to spaced repetition, OpenSen helps you build a personal language memory.

  1. Real-world situationAirport, café, class
  2. Dialog BuilderGenerate the conversation
  3. Chunk LibraryStore reusable patterns
  4. Recall practiceSpeak, don't just recognize
  5. Practice PlanFSRS schedule per chunk
  6. Automatic usageSay it without translating

Build your personal language memory

/04. Features

Tools for real learning.

Each module is built around chunks and real situations.

  • Situation Coverage

    Small talk, ordering, travel, school, and more.

  • Dialog Builder

    Memorization-ready dialogs from your own situation.

  • Chunk Library

    High-frequency native phrases with swap patterns.

  • Substitution Drills

    Keep the frame, swap the slot.

  • Practice Plan

    Recall that produces speech, scheduled with FSRS.

  • Anki Export

    Take your chunks anywhere.

/05. Product in action

From situation to language.

One lesson step from the app: a real situation becomes a sentence, a pattern, and a spoken answer.

  1. 1.Situation

    Illustration of a traveler asking an airport staff member for help

    At the Airport · Ask for help

  2. 2.Hear the sentence

    Could you help me find the station?

    SlowNatural

  3. 3.See the pattern

    • Could you help me
    • find
    • [PLACE]

  4. 4.Speak it

    Scored from your real speech, then scheduled for review.

/06. The experiment

I wasn't trying to build another Duolingo.

I wanted to test a hypothesis: can a learning system help people actually use the language they learn?

  1. 01

    Can chunking reduce the cognitive load of speaking?

  2. 02

    Can learners recall language faster when phrases are learned as units?

  3. 03

    Can AI generate useful, context-aware chunks instead of generic vocabulary exercises?

/07. What I learned

Progress, not perfection.

Some things worked, some didn't, and the project changed along the way.

  • What worked

    Learning phrases in context felt more natural than memorizing isolated vocabulary.

  • What didn't

    Generating lots of AI content quickly was less useful than curating a small set of high-quality examples.

  • What changed

    OpenSen moved from a dialogue generator toward a calm language studio: illustrated situations, curated sentences, and speaking practice.

/08. Next steps

Where OpenSen goes next.

A long-term exploration toward more personal and adaptive language learning.

Visit websiteInspiration videoConnect

  1. Shipped

    Learning loop

  2. Now

    Speaking practice

  3. Next

    Personal language graph

  4. Later

    Shareable packs

The long-term question

Can we build a language-learning system that adapts to the language a person actually needs to use?