seedance-food-asmr

Food close-ups and eating ASMR · 美食特写与吃播 ASMR

Cooking close-ups, mukbang and eating vlogs. They work when every beat shows one visible change in the food and one matching sound, and the dish stays the same dish from first frame to last.

This Skill carries one prompt structure distilled from 9 human-verified Seedance cases on goodcase.ai. It is a sibling of seedance-prompt-library (all templates in one Skill); install this one when you only want this kind of clip. Do not invent structure from general video-generation knowledge: follow the structure below and ground the draft in one anchor case from references/cases.md.

Use when

• Step-by-step cooking clips, glossy food ads with juice and steam, and handheld eating vlogs or spicy challenges where a person reacts to the food.
• Chinese: Step-by-step cooking shorts, cheese-pull juicy steaming food commercials, and handheld mukbang, spicy challenges — videos of people reacting to food.

Workflow

1. Collect inputs. Ask only for what is missing: subject or product, setting, duration, aspect ratio, whether reference images / audio exist, and which language the final prompt should be in.
2. Pick one anchor case. Read references/cases.md, choose the case closest to the request, and name it in the reply. One anchor, never an average of several.
3. Fill the structure block by block. Every block below must be present in the final prompt; an empty block is where prompts go vague.
   1. Opening line: duration, look (anime film, glossy commercial or handheld vlog) and the exact dish
   2. Style and light paragraph: macro close-ups, shallow depth of field, steam, warm light; for vlogs, the camera device and its flaws
   3. If a person is in it: identity lock, outfit and the room or street
   4. Timeline by step, raw ingredient to finished dish, one action and one texture change per segment
   5. Hero ending: the finished dish alone, slow push-in or arc, steam rising
   6. Audio: music style and tempo, then the cooking or eating sounds listed in the order they happen
   7. Negative tail: no text, UI or logos, this dish only, food and hands and utensils consistent
4. Apply the guidance and check the pitfalls (next two sections) before returning anything.
5. Return the finished prompt as one copy-pasteable block, the anchor case link, and a three-line checklist of what to verify in the generated video.

Guidance (from the cases)

• Cut the cooking into short named steps with timecodes. The katsudon case runs twelve segments of about two to three seconds each, titled Prepare Pork, Bread the Pork, Fry, Slice and so on, each with one action.
• Write the food's physical state, not just its name. The katsudon egg is set at the edges while the center remains glossy, slightly runny, and trembling, and the cut crust cracks naturally, revealing juicy white pork; the shengjianbao case has broth falling in long glossy strands and bottoms crisping into golden lace-like crusts.
• List the sounds in the order of the actions. The katsudon audio block asks to Synchronize realistic ASMR cooking sounds and names mallet, knife, sizzle, bubbling dashi, chopsticks and a ceramic clink, over city-pop at 110 to 120 BPM, ending on one wind-chime tone.
• Break eating into small steps and put the line after them. In the oden vlog she blows on it gently, then takes a bite, chews, looks at the camera, then says a short line; the spicy challenge lines up bowls from mildest to hottest so each bowl moves her reaction one step.
• Fence off the dish and the screen. The katsudon negative block says No unrelated ingredients or dishes. Katsudon only and bans subtitles, UI, logos and text overlays, then asks to keep food, hands and utensils consistent.

Pitfalls

• Hands and utensils warp in close-up, chopsticks multiply or the knife bends. Give each beat one utensil, keep the macro on the food, and add a line that hands and utensils stay consistent.
• The dish drifts halfway through, ingredients swap or a new dish appears. Name the dish, list its ingredients, and exclude unrelated food the way the katsudon case does.
• Too many shots squeezed into one paragraph. The shengjianbao case packs six shots into 15 seconds with no timecodes and jumps from eating back to cooking, so the model picks its own order; give each step at least two seconds and a timecode.
• Talking with a full mouth breaks the lip sync and the chewing. Let her bite, chew and swallow first, then say the line, the way the oden vlog spaces each bite and sentence.

Language

Reply in the user's language. The prompt itself can stay in English when that is what the user's Seedance workflow expects; ask if unclear. The Chinese version of this structure follows.

Chinese: Food close-ups and eating ASMR

Cooking close-ups, mukbang, and eating vlogs. It works because every beat makes the food visibly change, paired with a matching sound, and it's the same dish from start to finish.

Structure:

1. Opening line: duration, visual style (animated film, glossy commercial, or handheld vlog), and exactly which dish
2. Style and lighting section: macro close-ups, shallow depth of field, steam, warm light; for vlogs, write the shooting gear and its flaws
3. When someone is on camera: character lock, clothing, room or street
4. Write the timeline step by step, from raw ingredients to finished dish, one action plus one texture change per segment
5. Hero ending: the finished dish enters frame alone, slow push-in or orbit, steam rising
6. Audio: music style and tempo, then list the cooking or eating sounds in the order they happen
7. Exclusions and ending: no text, UI, or logos, only this one dish, food, hands, and utensils consistent throughout

Key points:

• Break the cooking into short steps with time codes, give each step a name. The pork cutlet rice bowl one has twelve segments total, each two to three seconds, with titles like Prepare Pork、Bread the Pork、Fry、Slice so each segment does only one thing.
• Write the physical state of the food; just naming the dish is useless. In the pork cutlet rice bowl, the egg edges set, the center remains glossy, slightly runny, and trembling, the cut-open crust cracks naturally, revealing juicy white pork; the pan-fried bun one writes the broth in long glossy strands dripping down, the bottom fried to golden lace-like crusts。
• List the sounds one by one in the order of the actions. The pork cutlet rice bowl's audio section writes Synchronize realistic ASMR cooking sounds, naming the meat tenderizer, the knife, the frying sizzle, the broth bubbling, chopsticks, and the light clink of the ceramic bowl, over 110 to 120 BPM city-pop, ending with a wind chime.
• Break the eating action into small steps, put the lines afterward. In the oden one she blows on it gently, then takes a bite, after chewing looks at the camera, then says a short line; the spicy challenge orders the bowls from least to most spicy, and with each bowl her reaction goes up a notch.
• Lock down both the dish and the visuals. The pork cutlet rice bowl's exclusion section writes No unrelated ingredients or dishes. Katsudon only, and also bans subtitles, UI, logos, and overlaid text, finally requiring the food, hands, and utensils to stay consistent throughout.

Common pitfalls:

• Hands and utensils deform in close-ups, chopsticks gain an extra one, the knife bends. Give only one utensil per beat, aim the macro at the food, and add a line that hands and utensils stay consistent.
• The dish changes halfway through, ingredients swap or another dish appears. Name the dish, list the ingredients, and exclude unrelated foods like the pork cutlet rice bowl one does.
• Cramming too many shots into one paragraph. The pan-fried bun one crammed six shots into 15 seconds, with no time codes, and jumped from eating back to cooking, leaving the order for the model to figure out; give each step at least two seconds and mark the time codes.
• Talking with food in the mouth, so lip-sync and chewing both break. Have her bite, chew, and swallow first, then say the line, spacing one bite per line like the oden one does.

Copy-ready lead-in / 可复制引导语

For users who would rather paste into a chat model than run this Skill, hand them this line above the structure:

I want a short food close-up video. [The dish is a bowl of tomato beef brisket noodles, filmed from slicing the tomatoes to serving.] [End on a steaming close-up of the finished bowl, with the sound of it bubbling on the stove.] Using the prompt template below, rewrite it into one ready-to-use Seedance video prompt for me:

I want to make a short food close-up video, [making a bowl of tomato beef brisket noodles, filming from cutting the tomatoes all the way to taking it out of the pot], [finally stopping on a close-up of the steaming finished dish, with bubbling stewing sounds]. Please help me rewrite it into a ready-to-use Seedance video prompt based on the following prompt template:

Notes

• references/cases.md is generated from data/ in LearnPrompt/awesome-seedance and refreshed daily; do not hand-edit an installed copy.
• Prompts in the reference file belong to their creators (linked on every card). Transfer the structure; do not present a close copy as original work.
• For the full library across every model, install the goodcase Skill from LearnPrompt/goodcase-lite.