Skip to content
VoqText

Languages

Interface languages and verified speech-recognition support, shown separately and honestly.

Interface languages

The product interface, emails and documents are available in these 14 languages. Arabic is fully right-to-left.

  • EnglishEnglish
  • EspañolSpanish
  • FrançaisFrench
  • DeutschGerman
  • ItalianoItalian
  • PortuguêsPortuguese (Portugal)
  • NederlandsDutch
  • العربيةArabic
  • Bahasa IndonesiaIndonesian
  • Tiếng ViệtVietnamese
  • ไทยThai
  • FilipinoFilipino
  • 日本語Japanese
  • 简体中文Chinese (Simplified)

Speech recognition languages

Spoken-language support depends on the configured speech provider. The table shows the 14 product languages against the provider’s verified list and its own published accuracy tiers; the provider supports additional languages beyond these.

Provider: elevenlabs · model scribe_v2Verified 2026-10-21
  • Speaker detection
  • Word-level timestamps
  • Automatic language detection
  • Audio event tags
LanguageTranscribeProvider accuracy tier
English (English)Excellent (≤ 5% WER)
Español (Spanish)Excellent (≤ 5% WER)
Français (French)Excellent (≤ 5% WER)
Deutsch (German)Excellent (≤ 5% WER)
Italiano (Italian)Excellent (≤ 5% WER)
Português (Portuguese (Portugal))Excellent (≤ 5% WER)
Nederlands (Dutch)Excellent (≤ 5% WER)
العربية (Arabic)Good (>10% to ≤20% WER)
Bahasa Indonesia (Indonesian)Excellent (≤ 5% WER)
Tiếng Việt (Vietnamese)Excellent (≤ 5% WER)
ไทย (Thai)Good (>10% to ≤20% WER)
Filipino (Filipino)High Accuracy (>5% to ≤10% WER)
日本語 (Japanese)Excellent (≤ 5% WER)
简体中文 (Chinese (Simplified))High Accuracy (>5% to ≤10% WER)

Tiers are the provider’s published word-error-rate ranges, reproduced as-is. Real accuracy varies with audio quality, accent and vocabulary; VoqText does not publish its own accuracy figure.

ElevenLabs Scribe (model scribe_v2). Limits and features verified against the official documentation on 2026-10-21: files up to 3 GB and 10 hours per request, speaker diarization up to 32 speakers, word-level timestamps, automatic language detection and optional audio-event tags. The provider publishes per-language word-error-rate tiers; results vary with audio quality, accent and vocabulary.

Availability by operation

  • Transcribe

    Available for the languages listed above.

  • Detect language

    Available when the provider supports automatic detection (shown above).

  • Translate

    Not available yet: transcript translation needs an approved translation provider. The interface itself is available in 14 languages.

  • Highlights and search

    Available for every transcribed language. Highlights and search results are quoted from the transcript, never written.

  • Subtitles

    Available for every transcribed language; CJK text is wrapped by characters.