Назад към всички

IMA Studio Music Generation

// Use for AI music generation via IMA Open API. Supports text_to_music with 3 models. IMPORTANT — Default model selection rule: always recommend the NEWEST and most POPULAR model, NOT the cheapest. Default: Suno (model_id=sonic, model_version=sonic-v5, 25pts) — latest engine, full-featured with custom

$ git log --oneline --stat
stars:1,933
forks:367
updated:March 4, 2026
SKILL.mdreadonly
SKILL.md Frontmatter
nameIMA Studio Music Generation
version1.2.2
categoryfile-generation
authorIMA Studio (imastudio.com)
keywordsimastudio, ai music, text_to_music, Suno, DouBao, GenBGM, GenSong
descriptionGenerate music from text with IMA Open API. Supports Suno (sonic), DouBao BGM (GenBGM), and DouBao Song (GenSong). Uses only api.imastudio.com.
requires[object Object]
persistence[object Object]

IMA Voice AI Creation

Capability

This skill generates music/audio from text prompts (text_to_music) through IMA Open API.

Supported model IDs:

  • sonic (Suno)
  • GenBGM (DouBao BGM)
  • GenSong (DouBao Song)

Network and Credential Transparency

  • API domain used: https://api.imastudio.com
  • Required key: IMA_API_KEY (environment variable)
  • The script does not call secondary upload domains.
  • The script does not read other skills' files.

Runtime Rules

  1. Always query /open/v1/product/list first.
  2. Resolve attribute_id, credit, and latest model_version from product list.
  3. Create task via /open/v1/tasks/create.
  4. Poll /open/v1/tasks/detail until completed or timeout.

Defaults and Timeouts

  • Task type is fixed to text_to_music.
  • Poll interval: 5 seconds.
  • Max poll wait: 8 minutes.
  • If --model-id is omitted, default model is sonic.

User Input Mapping

  • BGM / instrumental / 背景音乐 / 纯音乐 -> GenBGM
  • Song / lyrics / 人声 / 歌曲 -> sonic or GenSong
  • If unspecified -> default sonic

Script Invocation

Set key first:

export IMA_API_KEY="ima_your_key_here"

python3 {baseDir}/scripts/ima_voice_create.py \
  --model-id sonic \
  --prompt "upbeat lo-fi hip hop, 90 BPM, no vocals" \
  --output-json

List models:

python3 {baseDir}/scripts/ima_voice_create.py \
  --list-models

Error Handling Policy

  • Return user-friendly error summaries in plain language.
  • Include practical next step suggestions (retry, switch model, check API key/credits).
  • Do not expose raw internal payloads unless debugging is explicitly requested.

Expected Output

On success, return:

  • task id
  • result URL
  • model id/model name
  • credit used

If --output-json is enabled, parse JSON from script output.