How to Make Faceless YouTube Videos and UGC Ads with Higgsfield Skills in 2026

Two new skills are now available in Higgsfield MCP and Supercomputer. Type a topic or a product idea into the chat, and a finished video comes back without leaving the conversation. Faceless Content Factory turns a topic into an animated YouTube video up to 10 minutes long. UGC Factory turns a product idea into a creator-led ad with the same avatar carried through the whole clip. This guide covers how to set up MCP for both, a full workflow for each skill, and what it costs to run them.

What is Faceless Content Factory?

Faceless Content Factory takes a topic, a historical event, a how-something-works explainer, a story, and turns it into a full animated video for YouTube, up to 10 minutes long. You pick one of 13 ready visual styles, Stickman Cartoon, Editorial Motion Graphics, Claymotion, and more, then describe the topic. The skill writes the script, generates the narration, builds the animation, adds music, and drops in subtitles, all without you touching a script editor, a voice tool, or an animation program separately.

What is UGC Factory?

UGC Factory does the same kind of thing for product ads. Give it a product idea, a photo of the product, and a creator reference if you have one. It writes an ad concept and a script, then generates a vertical video. The same creator avatar appears consistently from the first frame to the last, across 6 ready templates: Unboxing UGC, Product Review UGC, SaaS UGC, Try-On UGC, Tutorial UGC, and Localization, which keeps the same voice and pacing when the ad needs to run in another language.

Both skills live inside Higgsfield MCP, which means they run right inside a normal ChatGPT or Claude conversation. They're also available directly in Supercomputer on Higgsfield.

• In Claude: Copy the connector URL → https://mcp.higgsfield.ai/mcp

  → Claude, then Customize → Connectors → name it Higgsfield, paste the URL → connect, sign in, and start.

• In ChatGPT: Find Higgsfield in the Plugins Directory → Add → sign in → connect and start creating.

You type what you want in plain language, the same way you'd ask either assistant a question, and a finished video comes back in the same chat.

Full Workflow: Step by Step

Workflow A: Faceless Content Factory via MCP (Claude or ChatGPT)

Step 1: Connect Higgsfield MCP if it isn't already. The same one-time setup covered above, for whichever assistant you're using.

Step 2: Pick a style and describe the topic. "Create a faceless history video in the Paper Diorama style about the Battle of Gaugamela, how Alexander the Great tore a gap in the battlefield Darius had flattened for himself, sent the king running, and rode into Babylon's open gates twenty days later." is enough to start. If you want more control over the result, you can write out the full prompt in detail instead, covering the exact beats, pacing, and tone you want the video to hit.

Step 3: Let the skill run. Claude or ChatGPT shows each stage, generating the video, adding narration, then montage and subtitles.

Step 4: Review the finished video. Script, narration, animation, music, and subtitles all arrive combined in one file.

Workflow B: UGC Factory via Supercomputer

Step 1: Open Supercomputer and attach references. A creator photo and a product photo both go into the same request.

Step 2: Describe the ad. "Create a complete 15-second product review UGC for my FIZZ POP bubblegum soda with this blonde creator: she talks to camera in a cozy kitchen, starts skeptical, pops the cap with a bottle opener on camera, takes a sip, shows the pink up close and ends with 'okay, I'm sold" is enough, the skill fills in the concept and script from there.

Step 3: Let the skill run. Supercomputer shows the stages as they happen, generating the video, adding audio, then montage and subtitles.

Step 4: Review the finished video. One complete vertical clip, the same avatar and product consistent from the first frame to the last.

Pricing

Both skills run on the same Higgsfield credit pricing as generating directly on the website, there's no extra charge for running them through MCP or Supercomputer instead. Here's what each workflow actually cost.

Which Faceless and UGC Styles to Start With

For Faceless Content Factory, the style you pick sets the tone before a single word of the topic gets written.

• Stickman Cartoon, for a topic that should feel simple and approachable, something a kid could follow along with.

• Editorial Motion Graphics, best suited to a topic that needs to read more like a documentary or a news breakdown, something with a bit of authority to it.

• Watercolor Chronicle, a good fit when the story calls for something softer and more illustrative, a historical moment or a quieter narrative.

• Fairy Tale & Myth, built specifically for a myth, a legend, or anything that plays like a fairy tale.

• Claymotion, worth reaching for when the story benefits from something tactile and playful, characters that feel handmade instead of flat.

The same logic applies to UGC Factory.

• Unboxing UGC, built around that reveal moment, useful when you're launching a new product and want to show it off in detail.

• Product Review UGC, for a creator-style product review or demonstration.

• Try-On UGC, the right choice when the product needs to be shown worn or used, a piece of clothing, a wearable, anything you'd want on camera instead of just held up.

What to Keep in Mind

• MCP/plugin setup is only needed when using the skills through a connected agent. Supercomputer workflows run directly inside Higgsfield and do not need MCP setup

• The 13 templates in Faceless Content Factory set the visual style, the topic itself only needs a single sentence.

• UGC Factory keeps one avatar consistent across the whole clip, the same reference carries through concept, script, and final video.

• UGC Factory covers 6 ready templates: Unboxing, Product Review, SaaS, Try-On, Tutorial, and Localization.

• Pricing matches the website exactly, running a skill through MCP or Supercomputer costs the same as generating the same output directly on Higgsfield.

• The UGC Factory skill inside Claude, ChatGPT, and Supercomputer runs on the same underlying product as the standalone UGC Factory on Higgsfield, just through a conversational, agent-driven front end instead of the full interface with Image Stage and keyframe control.

• The finished video runs well past the 15-second limit on a single MCP generation because the skill builds it from multiple generated segments assembled into one output, rather than a single continuous render.