AI Music & Songs · skill pubblica Open Skill
AI Music & Songs
AI Music & Songs è una Open Skill pubblica su Digen. Aprila per vedere cosa crea questo workflow.
Music
7 0 2
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4264620/2026-10-18/pgv5k93g.png
--- allowed-tools: - generate_music - generate_image --- # AI Music & Songs Create original AI-generated songs and music in the user's chosen language (Hindi, English, or others). Turn an idea, theme, mood, or genre into a full song with lyrics, or generate instrumental music. ## When to use - User wants to generate a song/music (Hindi, English, or other languages) from an idea, theme, mood, genre, or provided lyrics. - User wants to redo/refine a previously generated song, change its language, style, mood, or lyrics. - User wants a cover/album art image for a generated song. ## Steps 1. **Clarify the request** if needed: language (Hindi / English / other), genre/style (e.g. Bollywood, pop, rock, romantic, sad, upbeat, devotional, rap), mood, theme/topic, and whether they want vocals with lyrics or instrumental-only. If the user already gave enough detail, proceed. 2. **Write the lyrics** (if vocals are wanted) in the requested language. Match the genre and mood. For Hindi songs, write natural Hindi lyrics (Devanagari); for English, English lyrics. Structure them as verses, chorus, bridge as appropriate. 3. **Call `generate_music`** with: - `prompt`: describe the style, genre, mood, tempo, instruments, and language (e.g. "Upbeat Bollywood romantic pop song, female vocal, energetic, with tabla and strings, in Hindi"). - `lyrics`: the full lyrics text (omit for instrumental-only). - `model='music'`. - `name`: the song title. - `entity_type='music'`. 4. **Optionally generate a cover image** (if the user wants one): call `generate_image` with `model='t2i'`, `orientation='square'`, a prompt describing the cover art matching the song's theme/mood, `name` = song title + " cover", `entity_type='music_covers'`. 5. **Reply** in natural language summarizing the song (title, language, genre/mood) and mention the generated audio (and cover if made). Don't paste large URLs. ## Notes - Use `model='music'` for all music generation; never hard-code a physical model name. - If the user is unhappy with a result, ask what to change (language, style, mood, lyrics, tempo) and regenerate. - Report failures honestly; if generation fails, tell the user and offer to retry with adjusted parameters. - Politely decline requests for content that violates platform policy.