Singing Girl Video · Open Skill 公開スキル

Singing Girl Video

Singing Girl Video は Digen 上の公開 Open Skill です。開くと、このワークフローが何を作るか分かります。

@rudyofficial010

Music

18 6 2

https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4167672/2026-10-07/hzfneq52.jpg

# Singing Girl Video

## When to use
- The user provides a photo of a girl and wants a video of her singing a song/lyrics.
- Default lyrics: "thank you for sunshine, thank you for pain, thank you for joy, it's a beautiful day ay ay ay" (the user can supply different lyrics).
- Also use when the user wants to redo, tweak, or continue one of these singing-girl videos.

## Steps
1. Confirm the subject photo (the girl who will sing). If the user hasn't provided one, ask for it before proceeding.
2. Optionally call `recognize_image` on the photo to describe the subject (appearance, hair, clothing, setting) so the video prompt can keep her consistent.
3. Determine the lyrics: use the default lyrics above unless the user provides different ones.
4. Generate the sung vocal track with `generate_music`:
   - `prompt`: describe the song style (e.g. upbeat pop, warm and joyful, female vocal) and that it is sung by the girl in the photo.
   - `lyrics`: the chosen lyrics.
   - `name`: "singing track", `entity_type`: "music".
5. Generate the singing video with `generate_video`:
   - `model`: "ai2v" (image + audio → video, scene driven by the audio).
   - `input_urls`: [the subject photo].
   - `audio_url`: the music asset from step 4.
   - `orientation`: "portrait" (9:16, vertical for social media) unless the user asks for landscape.
   - `resolution`: "720P".
   - `prompt`: describe the girl singing the lyrics, lips moving in sync with the audio, joyful expression, tasteful music-video feel.
   - `name`: "singing video", `entity_type`: "final_videos".
6. Reply in natural language summarizing the result; offer to redo, tweak the lyrics/style, or continue.

## Notes
- Video generation does not support NSFW/adult content on any channel — keep the singing video tasteful and non-explicit even if the source photo is lingerie or suggestive.
- The `ai2v` channel requires both an image and a driving audio track; don't stuff the audio URL into `input_urls`.
- If the user wants to combine multiple clips into one video, use `mix_clips`.
- Report failures honestly and never fabricate URLs.