Speech-to-Text
Get Languages
Get the list of supported speech-to-text languages
GET
Request
Headers
string
required
Bearer token with your API key
Response
boolean
Always
true on 200 responsesstring
Status message (
"Languages retrieved successfully")object
Language catalog envelope
array
Ordered list of supported languages. The first entry is always
auto, representing auto-detect. Use the code value when submitting /stt requests.string
ISO 639-1 language code, or
"auto" for the auto-detect entrystring
Human-readable language name (e.g.
"English", "Hindi", "Arabic (MSA)")string
Same as
name for compatibility with clients that read the native keyboolean | array
false for European languages, or an array like ["en"] for Indic languages that support inline code-switching with the listed language(s)number
Target word error rate for this language at normal audio quality. Omitted for
auto.number
Total number of language entries (including the
auto option)object
Backend feature flags (e.g.
word_timestamps, speaker_diarization, srt_export, vtt_export, code_switching_indic_english, multi_language_per_request, max_languages_per_request, auto_language_detection, websocket_streaming, telephony_mulaw, continuous_mode)Notes
- The
autoentry is synthesized by this API and is not one of the backend’s native language codes. To submit a request with auto-detection, either omit thelanguagefield fromPOST /sttor explicitly passlanguage=auto. - Supported languages total 39 real languages + 1
autoentry. - Unsupported languages (explicitly rejected): Urdu (
ur), Japanese (ja), Korean (ko), Chinese (zh), Thai (th), Vietnamese (vi), Indonesian (id), Tagalog (tl), Swahili (sw), Turkish (tr), Persian (fa), Hebrew (he). - Arabic dialect tags (e.g.
ar-eg,ar-lv) are rejected — passlanguage=arfor best-effort MSA transcription.