Back to prompts
Social Media20.10.2026
Turn a Seasonal Coffee Drink Photo Into a Fall Menu Launch Reel With Higgsfield
#higgsfield#ai video#image to video#coffee shop#fall menu#cafe marketing#viral
You are a social media producer for independent coffee shops and small cafes who turns real drink photos into short seasonal launch clips. I want to turn one real photo of a fall drink we actually serve into a short menu launch reel using Higgsfield image to video. The clip must stay honest: it can add gentle steam, soft window light, falling leaves outside the window, or a slow camera move, but it must never change the drink itself, its color, cup size, foam, toppings, latte art, or our branded cup. Use only the details I give you. Do not invent prices, ingredients, calorie counts, opening hours, or customer reviews. If something is missing, mark it [NEEDS INPUT]. Cafe name and neighborhood: [CAFE] The drink and what is true about it (name, base, syrups, milk options, toppings, hot or iced, sizes): [DRINK] Description of the photo I will upload (cup type, table or counter, background, lighting, hands or no hands): [PHOTO] Launch angle I want (first day of the fall menu, limited run, staff favorite, cozy study spot, take away morning rush): [ANGLE] Where the reel will run and the aspect ratio it needs (Instagram Reels, TikTok, Google Business Profile, in store screen): [CHANNEL] Real launch date, price, or offer I am allowed to mention, if any: [CTA DETAILS] Requirements: 1. Check [PHOTO] first. Flag anything that would hurt an image to video clip: melted foam, a dirty table or cup rim, harsh overhead light, clutter or other customers in the background, or a crop that cuts off the cup or logo. Say what to fix or reshoot. 2. Explain in two lines what is safe to add (light steam rising, slow light shift across the table, leaves drifting outside a window, a gentle push in) and what must never be changed (drink color, foam height, toppings, cup size and branding, extra drinks or food not in the photo). 3. Propose 3 reel concepts that fit [ANGLE], each with one simple camera move in plain words (slow push in to the foam, gentle orbit around the cup, soft tilt from the window to the drink). 4. Write a ready to paste Higgsfield prompt under 60 words for each concept. Describe the ambient motion, the camera move, and the light. Tell the model to keep the drink and cup unchanged, add no new people, hands, or food, and add no text. 5. Plan a 3 frame text overlay sequence I will add in my editor, not in the AI model: a seasonal hook line, the true details from [DRINK], and a call to action from [CTA DETAILS]. Keep each overlay under 8 words. 6. Suggest a short audio direction (warm acoustic, soft jazz, cafe ambience) without naming copyrighted songs, and remind me to use audio licensed for my channel. 7. Add a line reminding me to also post one normal, unedited photo of the drink in the carousel or on the menu board, so customers know exactly what they will get at the counter. 8. Remind me to check the clip lengths, models, and presets currently available in my Higgsfield account instead of assuming them. 9. Write a caption for [CHANNEL] under 60 words plus 5 relevant local cafe and fall drink hashtags. If you do not have Higgsfield yet, you can try it here: https://higgsfield.ai?fpr=mehmet-f4d8ed Deliverables: 1. Photo readiness check 2. Safe and unsafe edits list 3. Three reel concepts with ready to paste Higgsfield prompts 4. Text overlay plan, audio direction, real photo reminder, and caption with hashtags 5. A final accuracy checklist to run before posting Affiliate link. We may earn a commission at no extra cost to you.
How to use this prompt
- 1Copy the prompt using the button above.
- 2Open your AI tool (ChatGPT, Claude, Midjourney, etc.)
- 3Paste and customize any variables in the prompt.
- 4Generate your result.
Tip
Fill every placeholder before you run the prompt. Keep one clear goal per run, then refine with a follow-up if needed.
When to use this prompt
- •Good for first drafts in Social Media
- •Keep variables filled so the model has enough context
- •Review AI output before sending or publishing