Create million-view miniature food videos using Claude AI and Higgsfield.

A step-by-step process for producing million-view miniature cooking videos (Miniature Cooking ASMR) entirely using AI with Claude AI, Higgsfield MCP protocol, and CapCut.

The niche of miniature cooking (ASMR) content is creating a huge sensation on short-video platforms, with videos racking up tens to hundreds of millions of views. The irresistible appeal of tiny pans, miniature ingredients, and realistic sizzling cooking sounds provides viewers with a highly satisfying visual and relaxing experience. This faceless content model boasts superior audience retention and RPM (Return on Video) revenue.

In this article, you will be guided through a comprehensive automation process from ideation and scripting to image creation and video animation, all within a single interface, thanks to the combination of the Claude AI model and the direct connection of Higgsfield AI's MCP (Model Context Protocol).

Sample concept: "Tiny Pancakes on a Mini Wood-fired Stove"

• Setting: A rustic vintage wooden kitchen corner with a tiny, red-hot earthenware stove. • Props: A mini cast iron pan (4cm diameter), a miniature whisk, a small milk bottle, and a tiny honey jar. • Food: A plate of three mini stacked pancakes, topped with a small melted butter cube and drizzled with glistening honey. • Movement & ASMR: The crackling sound of quail eggs, the sizzling of batter on the hot pan, the rising steam, and the skillful flipping of the pancakes. • Format: Vertical video (9:16, Shorts/Reels/TikTok), 6 scenes of approximately 20-30 seconds each.

Establish a connection between Higgsfield AI MCP and Claude AI.

Instead of constantly switching between multiple tool tabs, the MCP protocol allows Claude AI to directly call Higgsfield's image and video generation engine right within the chat window:

1. Access the Higgsfield AI platform, navigate to the MCP CLI section, and copy the connection command. 2. Open Claude AI > Go to Settings / Connectors > Click Add Connector. 3. Name it 'Higgsfield', paste the copied command, and confirm the connection.

The 4-step process for producing million-view miniature cooking videos.

Step 1: Master Prompt brainstorm ideas and choose a dish.

Copy the Master Prompt below and paste it into Claude AI to activate the stream of suggestions for 10 viral miniature food ideas:

You are a Creative Director and Viral Short-form Content Producer specializing in the Miniature Cooking ASMR niche on TikTok, YouTube Shorts, and Instagram Reels.

Please help me brainstorm video ideas following these steps:
1. Ask me about your target audience and desired food style (European, Asian, Street Food, Dessert).

2. After I provide this information, suggest 10 MINIATURE DISH IDEAS with the strongest visual and auditory ASMR effects (plenty of scenes like cracking eggs, pouring sauce, sizzling hot pans, melting butter).

3. Present this in a table format including: Dish Name | Visual Hook | Prominent Miniature Prop | Main ASMR Sound.

Let's start with the first question!

Step 2: Master Prompt creates a sequence of 6 identical keyframes via Higgsfield.

After selecting a suitable idea, paste the prompt below so Claude can automatically generate the scene breakdown and have Higgsfield create the shots:

Based on the chosen dish idea: [TAG THE DISH NAME]

Act as Art Director and follow this process:
1. Create a detailed script for 6 macro shots in the following order: Miniature ingredient preparation → Mixing/preparing the batter → Placing the pan on a mini stovetop → Pouring ingredients into the boiling pan → Flipping/cooking to a golden brown → Presenting the dish and drizzling with sauce.

2. Automatically write a standard English KEYFRAME IMAGE PROMPT for each shot (emphasize: extreme macro close-up, authentic tiny kitchen scale, human fingertips for scale, miniature cast-iron cookware, realistic food textures, warm cinematic lighting, photorealistic 8k, 9:16 vertical, no plastic look, no cartoon style).

3. Use the Higgsfield MCP tool to create images for Scene 1. After I review and confirm 'Yes', continue creating images for the following scenes to ensure consistency in the wooden kitchen background.

Step 3: Master Prompt animates the video scene by scene.

Once you have all 6 scene images, paste the following prompt so that Claude automatically generates motion commands and calls Higgsfield to render 6 smooth motion clips:

Now, convert the 6 still images you just created into 6 short video clips (4–5 seconds each) using Higgsfield's video engine:

1. For each scene, write a detailed VIDEO MOTION PROMPT in English: slow and steady macro camera movement, realistic hand movements, simulation of liquid physics (milk, butter, melting sauce), rising smoke (rising gentle steam), and sizzling oil bubbles.

2. Call the video render command for Scene 1 via Higgsfield MCP.

3. When the video is complete, wait for me to type 'Yes' to automatically create the next clip until all 6 scenes are complete.

Step 4: Post-process ASMR audio and merge timeline using CapCut

Download all 6 clips from Higgsfield to your device and open CapCut or Premiere Pro to complete the product:

• Arrange the 6 clips in the correct story order on the 9:16 timeline, trimming them to create a fast-paced rhythm (each scene 2.5 to 4 seconds long). • Add high-quality ASMR sound effects (Foley Sound Effects): the crisp cracking of eggshells, the sizzling of a hot pan, the clinking of tiny metal spoons against a ceramic plate, and the dripping of sauce. • Add a gentle acoustic or instrumental lofi track at -24dB volume for a warm background.

Checklist for pre-publication review

Lean manufacturing system

Kết nối Higgsfield MCP vào Claude AI
→ Master Prompt 1: Chọn món ăn & concept viral
→ Master Prompt 2: Tạo chuỗi 6 ảnh Keyframe đồng nhất bối cảnh
→ Master Prompt 3: Diễn hoạt 6 clip video kèm hiệu ứng khói/sôi
→ Tải clip về CapCut + Layering âm thanh ASMR Foley sống động
→ Xuất bản video ngắn triệu view trên đa nền tảng

The model combining Claude AI and Higgsfield MCP opens up the possibility of mass-producing high-quality miniature food videos without the need for expensive studios, macro cameras, or complex animation teams.