higgsfield-soul

Creates reusable Soul ID character profiles for consistent identity across Higgsfield image and video generations.

Updated Jul 15, 2026
One-click install
npx skills add https://github.com/executiveusa/buffer-blaster- --skill higgsfield-soul-executiveusa
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: higgsfield-soul
Source: https://github.com/executiveusa/buffer-blaster-/tree/main/skills/higgsfield/skills/higgsfield-soul
Command: npx skills add https://github.com/executiveusa/buffer-blaster- --skill higgsfield-soul-executiveusa

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Maintaining the same character's face and appearance across multiple AI-generated images and video shots is difficult, since models drift on identity when prompts mix appearance with motion. This Skill provides the workflows and prompt structures to lock character identity using Higgsfield Soul IDs. ## Core Features & Use Cases - Soul ID Reference Creation: Build strong reference images and multi-angle character sheets (front, 3/4, side, full body) on neutral grey backgrounds. - Identity vs. Motion Separation: Enforces a hard rule splitting every prompt into a static Identity Block and a temporal Motion Block to prevent face drift. - Character Anchor Blocks & State Tracking: Define ten per-shot attributes and maintain separate sheets per character state (injuries, outfit changes, transformations). - Soul Cast & Cinema Studio Integration: Configure AI actors with archetypes, budgets, and eras on Business/Team plans. - Use Case: A creator producing a multi-scene short film generates a 6-panel character sheet once, registers it as a Soul ID, then writes per-shot prompts that keep the protagonist's face identical across every scene, angle, and costume change. ## Quick Start Ask the AI to create a Soul ID character sheet for your protagonist and write identity-separated prompts for a multi-shot video sequence.

Frequently Asked Questions about higgsfield-soul

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I keep an AI character consistent across multiple video generations?

Create a Soul ID by uploading or generating a reference image, then reference that character in future prompts instead of re-describing the face. Use a multi-angle character sheet as the reference for stronger consistency across extreme angles and action shots.

What makes a good Soul ID reference image?

Use a front-facing or 3/4 angle portrait with even lighting, a neutral-to-slight expression, no blur, and a solo subject. Generate references with Soul 2.0 for fashion-forward looks, Nano Banana Pro for photorealistic sharpness, or Seedream 4.5 for style range.

Why does my character's face drift or morph during video generation?

Face drift happens when appearance descriptors are mixed with motion and camera language in one prompt. Split every prompt into an Identity Block (static descriptors only) and a Motion Block (camera, action, timing only) so the model does not re-read face features mid-motion.

How do I change a character's outfit without changing their face?

Build a split-panel sheet: a ghost-mannequin outfit panel on the left and a head-and-shoulders identity panel on the right declared as matching the input face. This gives the model one panel for wardrobe and one for identity, so the outfit change cannot pull the face with it.

When should I not use Soul ID?

Skip Soul ID for single one-off shots where consistency overhead is unnecessary, or when you want maximum creative variation, since locking a character constrains the style range of each generation.

Can I cast a real person's face into AI-generated video?

Yes, using the hybrid sheet method: generate the sheet normally, erase the head on body panels, and paste the person's real photo into the portrait panel. This requires the person's explicit permission, and source photos must stay out of published output.