weryai-music-generator（维雅音乐发生器）

WeryAI Music Generator

Generate WeryAI music with the official base skill for vocal-song and music-only workflows. In agent environments, default to an audio-first flow: use wait-music.js to submit and poll until final audio URLs are ready, keeping the wait bounded (10 minutes ceiling). Use submit-music.js plus asynchronous status polling only if the user explicitly asks for async behavior or a task ID. This API is mode-driven rather than model-driven, so the main choice is whether the user wants a VOCAL_SONG or ONLY_MUSIC result.

Example Prompts

• Submit an instrumental generation task and keep checking until you can show me the final audio.
• Make a vocal pop song with warm female vocals and custom lyrics, and give me the audio links.
• Turn this style brief into a piano-only meditation track and preview the request body first.
• Generate a song from this reference audio URL, but use async submit plus status polling instead of waiting in one command.
• Check the status of my WeryAI music generation task and tell me whether the results are ready.

Quick Summary

• Main jobs: music generation, song generation, instrumental generation, task status
• Main modes: ONLY_MUSIC, VOCAL_SONG
• Default mode: VOCAL_SONG
• Default options: genre=none, emotion=none, instrument=none, timbre=male
• Main trust signals: dry-run support, paid-run warning, explicit mode rules, media-source validation with auto upload for local reference_audio

Authentication and first-time setup

Before the first real generation run:

1. Create a WeryAI account.
2. Open the API key page at https://www.weryai.com/api/keys.
3. Create a new API key and copy the secret value.
4. Add it to the required environment variable WERYAI_API_KEY.
5. Make sure the WeryAI account has available balance or credits before paid generation.

OpenClaw-friendly setup

• This skill already declares WERYAI_API_KEY in metadata.openclaw.requires.env and primaryEnv.
• After installation, if the installer or runtime asks for required environment variables, paste the key into WERYAI_API_KEY.
• If you are configuring the runtime manually, export it before running commands:

export WERYAI_API_KEY="your_api_key_here"

Quick verification

Use one safe check before the first paid run:

node scripts/balance-music.js
node scripts/wait-music.js --json '{"type":"VOCAL_SONG","description":"A calm pop sketch","gender":"m"}' --dry-run

• balance-music.js confirms that the key is configured and the account is reachable.
• --dry-run confirms the request shape locally without spending credits.
• Real wait or submit commands still require available WeryAI balance.

Prerequisites

• WERYAI_API_KEY must be set before paid runs.
• Node.js >=18 is required because the runtime uses built-in fetch.
• reference_audio accepts http/https URLs or local/file sources. Local/non-http(s) sources are uploaded first via /v1/generation/upload-file.
• webhook_url must be a public http or https URL.
• Real submit and wait runs consume WeryAI credits.

Security And API Hosts

• Keep WERYAI_API_KEY secret and never write it into the repository.
• This skill supports directly passing local file paths. If a local file path is provided, the runtime will automatically upload the local file to the WeryAI server for processing.
• Optional override WERYAI_BASE_URL defaults to https://api.weryai.com. Only override it with a trusted host.
• Review scripts/ before production use if you need higher assurance.

Supported Intents

• Generic music request -> default to ONLY_MUSIC.
• Song, vocals, lyrics, singer, male/female voice -> use VOCAL_SONG.
• Reference audio request -> add reference_audio and require audio_name.
• Existing task -> check status instead of creating a new paid job.
• Account question -> check balance before running a paid task.

Default Configuration

Unless the user explicitly changes them, prefer:

• type: VOCAL_SONG
• timbre: male (gender: m)
• description: required unless styles is sufficient
• genre: none
• emotion: none
• instrument: none

Switch to ONLY_MUSIC when the user clearly asks for instrumental-only output, background music, soundtrack, ambience, BGM, or no vocals.

Supported Option Sets

Use these grouped options when the user wants structured controls instead of a freeform description.

Music Type

• VOCAL_SONG
• ONLY_MUSIC

Default: VOCAL_SONG

Genre Options

• POP
• ROCK
• RAP
• ELECTRONIC
• RNB
• JAZZ
• FOLK
• CLASSIC
• WORLD

Default: none

Emotion Options

• HAPPY
• RELAXED
• WARM
• ROMANTIC
• TOUCHING
• SAD
• LONELY
• DEPRESSED
• TENSE
• EXCITED
• EPIC
• MYSTERIOUS

Default: none

Instrument Options

• PIANO
• GUITAR
• BASS
• DRUMS
• STRINGS
• WIND
• SYNTHESIZER
• ELECTRONIC_SOUND
• FOLK_INSTRUMENT
• MIXED_ORCHESTRATION

Default: none

Timbre Options

• male
• female

Default: male

Mode And Parameter Guidance

Guide the user progressively instead of asking for every music field up front.

• If the user simply asks for music without clarifying, start from the default VOCAL_SONG mode.
• If the user clearly wants background music, soundtrack, ambience, meditation music, BGM, or instrumental audio, switch into ONLY_MUSIC.
• If the user asks for a song, vocals, lyrics, or singer gender, stay in VOCAL_SONG.
• If the user already provides lyrics, gender, or explicit vocal intent, use that directly and do not re-ask the same thing.
• If the user sounds unsure, translate their creative request into the closest valid mode and style tags rather than exposing raw API vocabulary too early.

Recommended guidance pattern

Use short operator-style guidance like this:

• General help: When the user asks "how to use this skill", DO NOT paste raw shell commands. Instead, explain the capabilities in natural language and give 2-3 prompt examples.
• Default run:

I can start with the default setup: vocal-song generation, male timbre, no fixed genre/emotion/instrument tags. If you want pure instrumental music, a different timbre, or specific tags, I can switch that before submission.

• Vocal switch:

If you want a song instead of instrumental music, tell me whether you want custom lyrics, the singer gender, or just a general vocal style, and I will route it as a vocal-song request.

• Parameter changes:

I can map your request into music settings. For example: pure instrumental background music -> ONLY_MUSIC, full song with lyrics -> VOCAL_SONG, pop + warm + piano -> genre/emotion/instrument tags, female voice -> timbre=female, style reference clip -> reference_audio + audio_name.

• Safety before paid runs:

Before I submit a paid task, I will show the final mode, key fields, and prompt so you can confirm them.

When to ask follow-up questions

Ask only for the smallest missing detail needed to submit safely.

• Ask whether the user wants vocals only if the request is ambiguous between song and instrumental.
• Ask for lyrics only when the user wants a vocal song and has not already provided lyric direction.
• Ask for gender / timbre only when the user cares about male vs female vocals. Otherwise keep the default male timbre.
• Ask for audio_name only when the user provides reference_audio.
• Do not ask every question if a valid request can already be formed from the user's brief.

Map user language to modes and fields

Use these common mappings:

• instrumental, background music, soundtrack, ambient track, piano piece, meditation music -> type: ONLY_MUSIC
• song, vocal song, singing, lyrics, male vocals, female vocals -> type: VOCAL_SONG
• use this reference track, match this demo, follow this audio sample -> reference_audio + audio_name
• make it pop, rock, jazz, epic, warm, piano, strings -> convert into supported styles keys from references/api-music.md
• genre, emotion, instrument, timbre -> accept direct option values and normalize them into API styles / gender

Confirmation block before submission

Before a paid run, show a concise confirmation block with the final payload choices.

Ready to generate

- type: `VOCAL_SONG`
- description: `A warm cinematic pop song with emotional build and polished production`
- timbre: `male`
- gender: `m`
- lyrics: `custom lyrics provided`
- genre: `POP`
- emotion: `WARM`
- instrument: `PIANO`
- reference_audio: `none`

Wait for confirmation or requested edits before running a paid submission.

Intent Routing

Use wait-music.js as the default path in agent environments to deliver final audio results in the same turn.

• If the request is instrumental or generic background music, route to ONLY_MUSIC.
• If the request is for a song, vocals, lyrics, singer gender, or timbre, route to VOCAL_SONG.
• If the user explicitly asks for asynchronous behavior or just a task ID, use submit-music.js.
• If the user already has taskId, use status-music.js instead of creating a new task.
• If the user wants to check account readiness first, use balance-music.js.

Preferred Commands

# Default audio-first flow
node scripts/wait-music.js --json '{"type":"ONLY_MUSIC","description":"A calm piano piece"}'

# Query task status
node scripts/status-music.js --task-id <task-id>

Workflow

1. Identify whether the user wants a vocal song or music-only output. If unclear, default to VOCAL_SONG.
2. Build the request with at least one of description or styles.
3. Normalize genre, emotion, instrument, and timbre into valid API fields when the user supplies structured options.
4. If the request includes reference_audio, also provide audio_name.
5. Use submit --dry-run or wait --dry-run when you need to verify the payload before spending credits.
6. Default to an audio-first execution:

- Run wait-music.js to submit the task and poll status until final audio URLs are ready or the maximum timeout of 10 minutes (600 seconds) is reached.

1. Use submit-music.js only when the user explicitly wants a task ID or async behavior without waiting.
2. Use status-music.js to re-check an existing task without creating a new paid task.

Input Rules

• type must be VOCAL_SONG or ONLY_MUSIC.
• At least one of description or styles must be present.
• For ONLY_MUSIC, do not send lyrics or gender.
• genre, emotion, and instrument option fields are normalized into styles.
• timbre is normalized into gender and is only valid for VOCAL_SONG.
• Only use style keys that exist in references/api-music.md.
• Do not invent undocumented fields.

Output

All commands print JSON to stdout. Successful result objects can include:

• taskId and taskStatus
• audios
• lyrics
• coverUrl
• balance
• requestSummary
• errorCode and errorMessage

See references/error-codes.md for the common error classes and recovery hints.

User-facing delivery requirement

• Always render the audios URLs directly to the user as clickable Markdown links. Do not just output the taskId.
• If multiple audio tracks are generated, render all of them using markdown links consecutively.
• Include a brief summary of the generation parameters used (e.g. from requestSummary or your initial payload).
• If timeout is reached before completion, return the taskId to the user and ask if they want you to check the status again. Do NOT show the raw node status command to the user; use it internally.

Definition Of Done

The task is done when:

• the request is validated locally without CLI-side errors,
• wait reaches a terminal result and returns user-visible audio URLs,
• or wait reaches timeout and returns the taskId plus an offer to check status again later,
• or submit (if explicitly requested) returns a task ID,
• or status returns a clear terminal or in-progress state for the target task,
• and the response tells the user whether lyrics and cover art are present.

Constraints

• Do not pretend this music API supports dynamic model discovery; it is mode-driven and documentation-based.
• Do not use local file paths for reference_audio.
• Do not re-run submit or wait casually, because each re-run creates a new paid task.
• Do not default to wait-music.js in agent environments for long-running generations.
• Do not broaden this skill into general audio editing, DAW automation, or file mastering workflows.

Re-run Behavior

• submit is not idempotent; re-running it creates a new paid task.
• wait is not idempotent for the same reason: it submits first, then polls.
• status is safe to re-run for the same task ID.

References

• Field rules and style enums: references/api-music.md
• Error handling guidance: references/error-codes.md