[{"data":1,"prerenderedAt":4},["ShallowReactive",2],{"article-doc:docs\u002Fapi-reference\u002F7.languages":3},"---\ntitle: Languages & Models\ndescription: Supported language and model identifiers.\n---\n## Languages\n\nThe service passes the `language` value straight through to the upstream STT endpoint. Use the language code expected by the upstream model.\n\nCommon examples:\n\n- `zh` — Chinese\n- `en` — English\n- `auto` — let the upstream auto-detect the language\n\nReal-time, [segment](\u002Fdocs\u002Fapi-reference\u002Fsegment-transcription), and [batch jobs](\u002Fdocs\u002Fapi-reference\u002Fbatch-jobs) all accept a `language` field. In the realtime gateway, update the language at any time with a `session.update` message.\n\n## Model and capability identifiers\n\n- `model` — offline STT model override (default `whisper-large-v3-turbo`)\n- `sttModel` \u002F `stt_model` — realtime STT model selection\n- `normalizerModel` \u002F `normalizer_model` — text normalizer model selection\n- `backend` — upstream backend selection\n\nNot all model fields are forwarded by the gateway; see [Session Configuration](\u002Fdocs\u002Frealtime\u002Fsession-configuration) for the current whitelist.\n\nSpecific identifier values depend on the upstream provider and are not contractually fixed by this gateway.\n\n## See also\n\n- [Live Translation](\u002Fdocs\u002Fconcepts\u002Flive-translation) — source language, target language, auto-detection, and code-switching\n- [Segment Transcription](\u002Fdocs\u002Fapi-reference\u002Fsegment-transcription) — sync multipart parameters\n- [Batch Jobs](\u002Fdocs\u002Fapi-reference\u002Fbatch-jobs) — async `audio_url` parameters\n- [Session Configuration](\u002Fdocs\u002Frealtime\u002Fsession-configuration) — realtime `session.update` fields\n",1790059118944]