Language detection
Identify language signals before choosing a downstream speech or text path.
Loading Shunya AI
Preparing the current workspace.
Language intelligence
Understand the language before choosing the operation.
SHUNYA AI treats language, script, normalization, translation, and transliteration as separate stages. The distinction helps teams see where a result came from and where the runtime still needs a governed artifact.
Processing stages
Deterministic foundations first.
Script and romanization
Keep script detection separate from language detection so Romanized and mixed-language input can be reviewed explicitly.
Normalization and entity protection
Prepare text while preserving names, numbers, dates, and other recognized entities for downstream work.
Translation and transliteration
The provider-neutral contract exists, but the current runtime does not have approved local translation or transliteration artifacts installed.
Language programme
Ten target languages, with status shown per operation.
| Language | Script | TTS | STT | Text transformation |
|---|---|---|---|---|
| English English | Latin | Available locally when the verified baseline probe passes | Runtime status reported at request time | Capability status reported at runtime |
| Hindi हिन्दी | Devanagari | Available locally when the verified baseline probe passes | Runtime status reported at request time | Capability status reported at runtime |
| Bengali বাংলা | Bengali | Verified SHUNYA voice pack required | Runtime status reported at request time | Capability status reported at runtime |
| Telugu తెలుగు | Telugu | Verified SHUNYA voice pack required | Runtime status reported at request time | Capability status reported at runtime |
| Marathi मराठी | Devanagari | Verified SHUNYA voice pack required | Runtime status reported at request time | Capability status reported at runtime |
| Tamil தமிழ் | Tamil | Verified SHUNYA voice pack required | Runtime status reported at request time | Capability status reported at runtime |
| Gujarati ગુજરાતી | Gujarati | Verified SHUNYA voice pack required | Runtime status reported at request time | Capability status reported at runtime |
| Kannada ಕನ್ನಡ | Kannada | Verified SHUNYA voice pack required | Runtime status reported at request time | Capability status reported at runtime |
| Malayalam മലയാളം | Malayalam | Verified SHUNYA voice pack required | Runtime status reported at request time | Capability status reported at runtime |
| Punjabi ਪੰਜਾਬੀ | Gurmukhi | Verified SHUNYA voice pack required | Runtime status reported at request time | Capability status reported at runtime |
This table describes the current programme and local verification boundary. It is not a claim that ten fully operational language stacks are installed.
Want to inspect the actual contracts?
The docs distinguish public API routes from the internal runtime and list unavailable states explicitly.