file *

歌曲文件，支持 WAV、MP3、FLAC、M4A。

callback_url

任务完成后的 HTTPS 回调地址。

curl -X POST "$BASE/accompaniment" -H "Authorization: Bearer $API_KEY" -F "file=@song.mp3"

返回：{ "task_id": "task_xxx", "status": "queued", "outputs": { "vocals": "...", "accompaniment": "..." } }

file *

待转谱的音频文件。

formats

可选，逗号分隔：midi,numbered,staff。

layout

可选：landscape 或 portrait。

curl -X POST "$BASE/transcribe" -H "Authorization: Bearer $API_KEY" -F "file=@song.wav" -F "formats=midi,numbered,staff"

返回：{ "task_id": "task_xxx", "outputs": { "midi": "...", "numbered_score": "...", "staff_score": "..." } }

file *

待分离的混音歌曲文件。

format

可选：wav 或 mp3，默认 wav。

curl -X POST "$BASE/separate" -H "Authorization: Bearer $API_KEY" -F "file=@song.flac" -F "format=wav"

返回：{ "task_id": "task_xxx", "outputs": { "vocals": "...", "piano": "...", "drums": "..." } }

reference

可选。原唱参考音频；不传时按待修人声自动生成音准曲线。

performance *

待修的人声或演唱音频。

curl -X POST "$BASE/tune" -H "Authorization: Bearer $API_KEY" -F "performance=@vocal.wav"
curl -X POST "$BASE/tune" -H "Authorization: Bearer $API_KEY" -F "reference=@original.wav" -F "performance=@vocal.wav"

返回：{ "task_id": "task_xxx", "status": "queued", "poll_url": "/api/public/tune/tasks/task_xxx" }；分析与自动重合成完成后，返回的 results.audio 可直接下载。

file *

待提升的音频文件。

lowpass

可选，true 时保留低通约束。

steps

可选，推理步数，默认 50。

curl -X POST "$BASE/quality" -H "Authorization: Bearer $API_KEY" -F "file=@source.mp3" -F "steps=50"

返回：{ "task_id": "task_xxx", "outputs": { "audio": ".../enhanced-48k.wav" } }

file *

待降噪的音频文件.

strength

可选, 4-30, 默认 12.

steps

可选, 推理步数, 默认 50.

curl -X POST "$BASE/denoise" -H "Authorization: Bearer $API_KEY" -F "file=@noisy.mp3" -F "strength=12"

返回:{ "task_id": "task_xxx", "status": "queued", "results": { "audio": ".../download" } }

file *

歌曲文件，系统会先进行人声处理。

formats

可选：lrc,ttml,srt,txt。

language

可选，歌词语言代码，如 zh。

curl -X POST "$BASE/lyrics/generate" -H "Authorization: Bearer $API_KEY" -F "file=@song.mp3" -F "formats=lrc,ttml,srt"

返回：{ "task_id": "task_xxx", "outputs": { "ttml": "...", "lrc": "...", "srt": "..." } }

lyrics *

待校对的歌词文本或带时间轴歌词。

reference

可选，用于对齐的原始校对稿。

script

可选：simplified 或 traditional。

curl -X POST "$BASE/lyrics/proofread" -H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" -d '{"lyrics":"...","script":"simplified"}'

返回：{ "lyrics": "...", "changes": [{ "from": "", "to": "" }] }

audio_url *

可公开访问的音频地址。

effect

可选：grid,mist,partita,scope 等。

theme

可选：dark 或 light。

curl -X POST "$BASE/player/spectrum" -H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" -d '{"audio_url":"https://.../song.mp3","effect":"grid"}'

返回：{ "embed_url": "...", "config": { "effect": "grid" } }

audio_url *

可公开访问的音频地址。

lyrics_url *

TTML 或 LRC 歌词文件地址。

lyric_mode

可选：classic,cadenza,fume,partita。

curl -X POST "$BASE/player/floating-lyrics" -H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" -d '{"audio_url":"https://...","lyrics_url":"https://.../song.ttml"}'

返回：{ "embed_url": "...", "config": { "lyric_mode": "classic" } }

audio_url *

可公开访问的音频地址。

lyrics_url *

逐字 TTML 歌词地址。

mode

可选：curtain 或 karaoke。

curl -X POST "$BASE/player/curtain" -H "Authorization: Bearer $API_KEY" -H "Content-Type: application/json" -d '{"audio_url":"https://...","lyrics_url":"https://...","mode":"karaoke"}'

返回：{ "embed_url": "...", "config": { "mode": "karaoke" } }

file *

待分析的歌曲文件。

prompt_style

可选：generation,arrangement,lyrics。

callback_url

可选，任务完成后的 HTTPS 回调地址。

curl -X POST "$BASE/reverse-analysis" -H "Authorization: Bearer $API_KEY" -F "file=@song.mp3" -F "prompt_style=generation"

返回：{ "task_id": "task_xxx", "analysis": { "bpm": 120, "key": "C major", "segments": [] } }