About Kotoba Engine
Why this exists
Dictionaries and translation tools are great at telling you what a Japanese sentence means. They're much worse at telling you how it lands — whether it sounds warm or cold, whether a native speaker would actually say it that way, whether it's the kind of thing you'd text a friend or say to your boss. That gap — between the dictionary definition and the lived, social meaning — is what Kotoba Engine is for. Real sentences, annotated by real learners and speakers, word by word, for exactly the nuance a dictionary entry can't carry.
Where the AI fits, and where it doesn't
Every entry gets an automatic pragmatic read from an AI model, and every word gets a computed reading from a real morphological tokenizer — that's useful scaffolding, not the final answer. You'll see a note on AI-generated reads reminding you of exactly that: it's a starting point, and the actual point of this board is people who know better adding their own annotation on top of it. If an AI read looks off to you, that's not a bug to report — that's the invitation.
Who's building this
Kotoba Engine is an independent, early-stage project built by Sushil Raj (CSE, SRMIST Kattankulathur) — not a company, not a funded startup, just something built because the gap it fills felt real. That also means it's small right now: a young community, a growing set of entries, still finding its shape. If you've found this and you're reading this page, you're early — which mostly means your first few posts and annotations will matter more here than they would on a board with ten thousand members already.
The source is open
The whole codebase is public — see the “Source” link in the footer. If something's broken, missing, or you just want to see how a particular feature works, it's all there to read.