ByteDance Seed Speech Text to Speech
/v1/runSeed Speech TTS 2.0 is ByteDance's text-to-speech AI model. Generate human-like voiceovers with expressive intonation, multilingual support, and customizable voice characteristics.
Request body
Submit an async generation request. The model field selects the model; other fields are model-specific input parameters.
Model identifier. Set to bytedance/seed-speech/tts/2.0.
Default: bytedance/seed-speech/tts/2.0
The text to synthesize into speech.
Output audio format. 'mp3' returns MP3 audio; 'opus' returns Opus audio in an Ogg container.
Allowed values: mp3, opus
Default: mp3
Sample rate of the output audio in Hz.
Allowed values: 8000, 16000, 22050, 24000, 32000, 44100, 48000
Default: 24000
Volume. 1.0 is normal volume, 0.5 is half, 2.0 is double.
Range: 0.5 to 2
Default: 1
Voice to use for synthesis. The preset name encodes the voice and its supported language codes. 'mixed_en_zh' means the voice can seamlessly blend English and Chinese; separate codes (e.g. 'en_zh') mean the voice supports each language independently.
Allowed values: vivi_mixed_en_zh_ja_es_id, mindy_en_es_id_pt_zh, stokie_en, dacey_en, tim_en, kian_en_zh, cedric_en_zh, sophie_en_zh, jean_en_zh, magnus_en_zh, mabel_en_zh, nadia_en_zh, opal_en_zh, pearl_en_zh, quentin_en_zh, vienna_mixed_en_zh, alina_mixed_en_zh, corinne_mixed_en_zh, esther_mixed_en_zh, freya_mixed_en_zh, gigi_mixed_en_zh, holly_mixed_en_zh, lyla_mixed_en_zh, daisy_mixed_en_zh, tracy_es_zh, jess_ja_es_id_pt_en_zh, pinky_es_ko_mixed_en_zh, sweety_ja_es, sandy_es_mixed_en_zh, sven_de, minimi_ja, usseau_fr, felipe_es, han_id, martins_pt, enzo_it, shane_ko, bonnie_zh, felix_zh, celeste_zh, monkey_king_zh
Default: stokie_en
Voice pitch shift in semitones. 0 is normal pitch, -12 lowers by one octave, 12 raises by one octave.
Range: -12 to 12
Default: 0
Force the text to be read as a single language, disabling automatic language detection. Leave unset for automatic detection (including seamless Chinese/English mixing on bilingual voices). Codes: zh (Chinese), en (English), ja (Japanese), es-mx (Mexican Spanish), id (Indonesian), pt-br (Brazilian Portuguese), ko (Korean), it (Italian), de (German), fr (French).
Allowed values: zh, en, ja, es-mx, id, pt-br, ko, it, de, fr
Optional natural-language instruction that steers the delivery (tone, emotion, pace, volume), e.g. 'Speak in a cheerful tone' or 'Could you speak a bit slower?'. It is not spoken aloud and does not affect billing.
Speech speed. 1.0 is normal speed, 0.5 is half speed, 2.0 is double speed.
Range: 0.5 to 2
Default: 1
Response Schema
The submit endpoint returns a run response. If its status is pending or running, poll GET /v1/run/{id} with the returned opaque ID until it reaches a terminal state.
Opaque SandBase run identifier. Use it exactly as returned; no prefix is guaranteed.
Current public run status.
Allowed values: pending, running, completed, failed, timeout
Public SandBase model name used for this run.
Present only for completed runs. Each object is capability-specific; inspect the selected model schema for its fields.
Present only for failed or timeout runs. Contains a public error type and sanitized message.
Stable public error category.
Sanitized error message safe to show to clients.
Usage details when available.
Model capabilities
Capabilities declared by the model registry.
Default: text-to-speech
Execution mode declared by the model registry.
Default: async