Language support, without the asterisk hiding the truth.
SaidClear lets you pin one of ten spoken languages for each capture and always keeps source text beside translations. Recognition quality depends on the configured browser or transcription provider.
The capability matrix
Choose one spoken language per capture. Recognition and translation depend on the configured provider; dense within-sentence code-switching is not yet verified.
Swipe sideways to compare every capability.
| Language | Browser capture | Translate to English | Localized interface | Dense code-switching |
|---|---|---|---|---|
| EnglishEnglish | Provider dependent | Not available | Available | Not available |
| SpanishEspañol | Provider dependent | Provider dependent | Available | Not available |
| Chinese中文 | Provider dependent | Provider dependent | Available | Not available |
| Hindiहिन्दी | Provider dependent | Provider dependent | Available | Not available |
| FrenchFrançais | Provider dependent | Provider dependent | Available | Not available |
| GermanDeutsch | Provider dependent | Provider dependent | Available | Not available |
| Japanese日本語 | Provider dependent | Provider dependent | Available | Not available |
| PortuguesePortuguês | Provider dependent | Provider dependent | Available | Not available |
| ItalianItaliano | Provider dependent | Provider dependent | Available | Not available |
| Arabicالعربية | Provider dependent | Provider dependent | Available | Not available |
What stays consistent in every language
- Original first
- Translations are additive. They never replace the words captured as source evidence.
- One pinned language
- Each capture records the language selected at the start. This avoids pretending that dense mixed-language recognition is solved.
- Correct direction
- Arabic source and translated content render with right-to-left semantics; the English interface remains left-to-right.