How to Generate LinkedIn Carousels and Visual Content on Autopilot

Claude Co-work could do almost everything except generate images and videos. The sandbox it runs in blocked all outside image and video generation APIs - Midjourney, Runway, OpenAI's image tools, Gemini. That meant every time you needed a visual asset for a LinkedIn post or carousel, you had to stop, switch tools, and come back. Higgsfield's connector fixes this. Here's what that unlocks for LinkedIn content.

Why this matters for LinkedIn GTM content

The workflow that wasn't possible before

What Higgsfield adds that other tools don't

Four steps to get the full system running. The setup takes about 15-20 minutes and you only do it once per content project or client.

Step 1: Create a dedicated working folder

One folder per content project

Step 2: Create a project with a CLAUDE.md file

Train Claude on your brand once

Step 3: Connect the Higgsfield connector

One-time connector setup

Step 4: Install the skills pack

Load the skills as a plugin

The carousel skill generates a full LinkedIn carousel - copy, slide structure, and on-brand visuals - from a single prompt. This is the highest-leverage use case for B2B LinkedIn content because carousels consistently outperform single-image posts on engagement and saves.

How to run it

The command

What it produces

The full output

Customising the output

How to make it match your style exactly

Iterating without wasting tokens

The real leverage from this setup isn't running it manually each time. It's scheduling it so Claude generates your LinkedIn content while you're not at your desk. You come back to finished carousels ready to review and post.

How scheduled tasks work in Co-work

Setting up a recurring content task

The weekly LinkedIn content production system

What this looks like in practice for a GTM founder

Turning good outputs into permanent skills

Save any style you like as a skill