A real Next.js dashboard foundation for the Kifuliiru phonetics work.
Included
- Overview dashboard
- Interactive phonetics generator
- KP practical notation + IPA candidate output
- Rule registry with Verified vs Review status
- Lexicon workspace placeholder
- Native-speaker validation queue
- Settings / architecture page
POST /api/generate API- Responsive desktop/mobile dashboard
- Dark/light system theme
- Phonetics engine isolated in
lib/kp.ts
Current established rules
- Double vowels (
aa ee ii oo uu) are one continued vowel, not two.
b is soft everywhere in Kifuliiru.
- KP practical notation:
bh
- Current IPA candidate:
[β̞]
j is soft like French jour.
- Current IPA candidate:
[ʒ]
- Spoken accent/prosody matters, but is intentionally not implemented until documented with native-speaker examples.
Mappings for sh, ny, ng, and ch remain candidates and are visibly marked for review.
Run
API
POST /api/generate
Content-Type: application/json
{"text":"niyiji"}
Production roadmap
The UI is intentionally structured around the pieces needed for a real phonetics product:
- persistent Supabase lexicon/rules/reviews
- authentication and reviewer roles
- rule versioning / KP releases
- native-speaker validation history
- lexical exceptions
- test corpus and regression testing
- confidence/evidence metadata
- spoken accent/prosody model
- KSE adapter and audio comparison
- import from Kifuliiru Dictionary / Tabula
- KEd integration
- analytics and generation history
The rule engine stays provider-independent: KP → KPG → KSE.
