Synthetic demonstration data only. No Tribal, student, or contributor data is used here. Stand-in language content is used, and no Native language is represented. This environment is not a FedRAMP-authorized or production system.

Language support

Script rendering, instructional design and speech services are kept as separate layers. A language without a standardized written form can still be taught through audio-first activities, and no orthography is invented.

Layer 1

Script and display

  • Full Unicode (utf8mb4) storage, normalized to NFC on save Live and demonstrable
  • Per-tenant character keyboard for authors and learners Live and demonstrable
  • Open fonts with broad coverage (Noto), with a tenant font if one is needed Configurable with existing product
  • Diacritic-aware and diacritic-insensitive search Live and demonstrable
  • Operating-system input methods or keyboard layouts for a specific orthography Requires integration
  • Language-specific collation (sort order) Future / pilot development

Layer 2

Content and instruction

  • Audio-first patterns: listen & select, listen & sequence, record & submit Live and demonstrable
  • “Written form approved” flag per item. Without it, the item stays audio-only Live and demonstrable
  • Learner interface strings localized per tenant, so no English prompts are needed Live and demonstrable
  • BCP 47 language tags with private-use subtags for community varieties Live and demonstrable
  • Phonics / sound-pattern sequences designed with the community’s speakers and a linguist Future / pilot development

Layer 3

Speech and language services

  • Neutral adapter interface for speech-to-text, text-to-speech and pronunciation analysis Requires integration
  • Human review is the default. Recordings are never auto-scored when no service has been validated Live and demonstrable
  • Safe degradation when an adapter is down: nothing is lost and work goes to people Live and demonstrable

Integration adapters →

Unicode lab

Type or use this tenant’s keyboard. The lab shows the normalized form and each code point. A decomposed sequence (base letter + combining mark) and its precomposed form are stored identically after NFC.

Diacritic-insensitive search over approved written forms

Rendering test (individual characters and combining marks, not words in any language)

ą́ ę̄ į̨́ ǭ̨ ų́ ł ŋ ʔ ʼ ə̀ ɨ̄ x̌ ⁿ ʷ [ˈʔa.ɬə ʃt͡ʃ ɣ ɾ ʎ] مرحبا Γειά σου 你好