How to Generate Images and Videos in Claude with Higsfield AI. Claude can now generate images and videos directly inside chat -- and the setup takes less than a minute. In this video, I'll walk you through how to connect Claude to Higsfield, show you exactly how it works with real examples, and help you start creating media without leaving your conversation.

Try Higgsfield MCP here: https://higgsfield.ai/s/mcp-teacherstech-xbTaOD

Higsfield MCP URL: https://mcp.higgsfield.ai/mcp

Higsfield is a creative AI platform that just launched an MCP connector for Claude. Once you turn it on, you can ask Claude to generate still images and video clips right inside your chat. Claude handles the connection in the background, so there's no coding or technical setup involved. I'll cover how to install the connector step by step, how to generate your first image, how to pick specific models like GPT Image 2 or Cance 2.0 for different results, how to refine and iterate on your creations, and how to generate cinematic video clips with detailed prompts. I also walk through a real-world example where I use Claude and Higsfield together to plan and produce social media content for a small business.

Timestamps:
0:00 Intro
0:30 What is Higsfield?
0:48 How the MCP connector works
1:16 Available image and video models
1:42 Setting up the connector step by step
2:52 Generating your first image
3:37 Choosing a specific model (GPT Image 2)
4:52 Refining and iterating on images
5:35 Generating video with Cance 2.0
6:01 Tips for writing better video prompts
7:25 Comparing video models side by side
9:22 Real-world workflow -- social media for a coffee shop
11:19 Honest thoughts and limitations

PROMPTS USED IN THIS VIDEO

1. Default image generation (routes to Nano Banana):

Generate an image of a small bakery storefront on a quiet street in autumn. Warm light coming from inside. Photographic style.

2. Specifying GPT Image 2 by name:

Using the Higgsfield MCP, generate the same bakery image but use GPT Image 2 this time.

3. Refining the image (no MCP prefix needed -- Claude carries it forward):

Take the GPT Image 2 version. Make it dusk instead of midday, and add a person reading on the bench outside.

4. Cinematic 10-second video with Seedance 2.0:

Using the Higgsfield MCP, generate a ten-second cinematic video clip with Seedance 2.0. Scene: a small independent bakery storefront on a quiet cobblestone street in autumn at golden hour dusk. Warm amber light spills through the front windows onto the wet cobblestones below. A wooden hand-painted sign hangs above the door, gently swaying in a light breeze. Autumn leaves in shades of orange, red, and yellow drift slowly past the storefront throughout the entire clip, catching the light. A young woman in a wool coat sits on a bench just outside the bakery reading a book, occasionally glancing up at the falling leaves. Camera movement: start with a slow wide establishing shot from across the street showing the full storefront, then push in smoothly over four seconds toward the bakery, then transition to a low angle medium shot of the bench and the bakery door, and finish with a close-up detail shot of leaves falling past the warm glowing window light. Three distinct camera angles in total. Cinematic depth of field, gentle natural motion, no text overlays, no people walking past, soft ambient atmosphere, golden hour color palette, photorealistic, shallow focus on the foreground subject in each shot.

5. Planning workflow demo:

I run a small bakery in a mountain town. Help me plan one Instagram post for this weekend. I want a still image and a short video clip. Keep the tone warm and inviting. Plan it first, then generate the assets through the Higgsfield MCP once we agree on the plan.


Subscribe: https://www.youtube.com/@TeachersTech?sub_confirmation=1
Website: https://www.teachers.tech
Instagram: https://www.instagram.com/teacherstechlab
LinkedIn: https://www.linkedin.com/in/teacherstech