Your hot spring boyfriend · skill pubblica Open Skill
Your hot spring boyfriend
Your hot spring boyfriend è una Open Skill pubblica su Digen. Aprila per vedere cosa crea questo workflow.
Anime
4 0 2
https://d7qlwi7tl2u27.cloudfront.net/skill-covers/prod/4156339/2026-09-22/7cr62f4n.jpg
# Onsen Handsome Man Interactive Short
## When to use
- The user wants to create a short anime-style "onsen handsome man" interactive video (scene animation + voiceover dialogue + BGM mix) from a one-liner or a scene.
- The user wants to reuse the same handsome man character (after uploading a reference image, always use the same character).
## Steps
1. First confirm the one-liner / scene the user provides; check whether a handsome man character reference image was uploaded in this session. If there is no reference image, ask whether the user wants to upload one, or directly generate a brand-new anime-style handsome man character.
2. Automatically write a short ~10-second story with 1–2 lines of dialogue (Chinese), deciding the scene action and the lines.
3. Generate the scene image (anime style, handsome man in an onsen / hot spring):
- With a character reference image: `generate_image(model='i2i.hd', input_urls=[reference image], prompt=<scene description>)`, `orientation='portrait'`, `name='scene'`, `entity_type='storyboard_images'`.
- Without a reference image: `generate_image(model='t2i', prompt=<scene description>)`, same parameters.
4. Generate the voiceover dialogue: `generate_tts(model='tts.design', text=<lines>, voice_instruction=<English male voice description>)`, `name='voiceover'`, `entity_type='voiceovers'`.
5. By default do not generate BGM, only the voiceover dialogue; generate BGM only when the user explicitly asks for it: `generate_music(prompt=<atmosphere description>)`, `name='bgm'`, `entity_type='music'`.
6. Generate the video: `generate_video(model='ref2v', input_urls=[character reference image or scene image], prompt=<animation description>, orientation='portrait')`, `name='clip'`, `entity_type='storyboard_videos'`.
7. Mix: by default mix in only the voiceover dialogue `mix_clips(clips=[{video: <video>, audio: <voiceover>}], orientation='portrait')`; if the user asks for BGM, use `mix_clips(clips=[{video: <video>, audios: [<voiceover>, <BGM>]}], orientation='portrait')`. `name='final'`, `entity_type='final_videos'`.
8. Summarize the finished video in natural language and tell the user they can keep adjusting the story, lines, or character.
## Notes
- By default do not generate BGM, only the voiceover dialogue; add BGM only when the user explicitly asks for it.
- Keep the audio clean: only the voiceover, BGM (background music), and a small amount of quiet water sound; no other sounds.
- The character's facial expressions should be lively and change naturally with the story and lines; do not hold a single expression throughout.
- When BGM is included, the voiceover must always be 15–20dB louder than the BGM; in mix_clips, lower the BGM volume to 0.15–0.25 and keep the voiceover volume at 1.0.
- Within a single video the background stays fixed; only the character moves.
- The platform generates videos up to 10 seconds, so all character dialogue voiceover must not exceed 10 seconds; if the user's requested dialogue audio exceeds 10 seconds, remind the user and automatically intelligently shorten the dialogue to within 10 seconds.
- Unless the user specifies otherwise, characters default to shirtless with only a towel wrapped around the lower body.
- If the character has lines, the video duration must be longer than the time it takes to finish the voiceover, so the picture is still playing when the lines end.
- Video generation does not support NSFW / adult content; politely decline such requests.
- If the user wants a permanently fixed character, they can upload a reference image and bind it as a skill preset asset, then reuse it automatically every time.
- Intermediate products (scene image, voiceover, BGM, single video clip) will be visible on the canvas; this is normal.