audio-song-generate

Audio Song Generate

Overview

Lyrics-driven full song generation (including vocals, complete structure of verses/bridge/chorus), multi-style creation (pop/classical/electronic/rock, etc.). Also supports strong-rhythm / full-arrangement instrumental music, with arrangement capability superior to general-purpose instrumental music tools.

API Mapping

• Song and strong-rhythm instrumental music generation:audio_cashew_songgenerate

Dependencies

• meitu-cli: >=2.0.6
• Credentials：CONFIG AKSK → meitu tools update; EXEC AKSK → actual execution (see root CONFIG.md）
• Environment variables：MEITU_OPENAPI_TOOL_TASK_MODE=command

Path alias:$VISUAL = {OPENCLAW_HOME}/workspace/visual/

Core Workflow

Preflight → Execute → Deliver

Preflight

1. meitu --version ≥ 2.0.6
2. Already ran with CONFIG AKSK meitu tools update
3. Current AKSK = EXEC,MEITU_OPENAPI_TOOL_TASK_MODE=command
4. output_dir：openclaw.yaml → ./output/ ｜else → $VISUAL/output/audio-song-generate/；mkdir -p

Execute

Trigger signals / routing rules

Core dimensions:Vocal/arrangement scenario hit + complete parameters。

Scenario	Determination keywords	Routing
Full song with vocals	write a song, with vocals, with lyrics, sing a song, brand song	audio_cashew_songgenerate
Strong-rhythm / fully arranged instrumental music	strong rhythm, strong variation, reversed melody, full arrangement feel	audio_cashew_songgenerate

Single API tool: call when parameters are complete, retry once on failure then return error; no cross-tool fallback. When style and lyrics conflict, prompt(style) prevails.

Parameter definitions

Parameter	Type	Required	Range	Default	Description
prompt	STRING	Yes	--	--	Song style description. If missing → prompt "please describe the song style you want"
prompt_lyrics	STRING	Yes	--	--	Lyrics content. If missing → prompt "please provide lyrics content" (for strong-rhythm instrumental music scenarios, can be filled with an empty string or stylistic description)
model	STRING	No	auto / mureka-6 / mureka-7.5 / mureka-o1	auto	Generation model selection, illegal values fall back to auto

audio_cashew_songgenerate Use prompt + prompt_lyrics + model. Vocal quality is affected by multilingual pronunciation.

Tool invocation

meitu audio-song-generate --prompt "<style>" --prompt_lyrics "<lyrics>" [--model auto] --json --download-dir {output_dir} --skill_name skill_audio-song-generate

Error degradation

Scenario	Handling method
prompt missing	Prompt "please describe the song style you want", do not call API
prompt_lyrics missing (vocal scenario)	Prompt "please provide lyrics content", do not call API
model specified a nonexistent value	Fall back to auto
audio_cashew_songgenerate invocation failed	Retry once, if still failing return error
prompt and prompt_lyrics style conflict	Use prompt as the standard, inform the user that the lyrics will be arranged in the specified style
User only wants lyrics, not melody	Refuse, not within the scope of this tool
Content compliance interception	Directly return compliance prompt, do not retry

Deliver

• Use output_dir resolved by Preflight
• From downloaded_files[0].saved_path read the downloaded audio file path
• mv {downloaded_files[0].saved_path} {output_dir}/{YYYY-MM-DD}_{descriptive}_audio-song-generate.mp3

Output

• Format：MP3
• Location: project → ./output/, one-time → $VISUAL/output/audio-song-generate/

Baseline Task ID

See references/task-id-baseline.md corresponding row.