higgsfield-soul-id

Train a face-faithful Soul identity model for Higgsfield image and video generation.

Updated Aug 6, 2026
One-click install
npx skills add https://github.com/ferrarifankid04/ai-skills-public --skill higgsfield-soul-id-ferrarifankid04
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: higgsfield-soul-id
Source: https://github.com/ferrarifankid04/ai-skills-public/tree/main/claude-code/skills/higgsfield-soul-id
Command: npx skills add https://github.com/ferrarifankid04/ai-skills-public --skill higgsfield-soul-id-ferrarifankid04

SYSTEM DOCUMENTATION & REQUIREMENTS

💡 This Skill includes references (resource) components.

What problem does it solve? Creating consistent, identity-faithful images and videos of a specific person with AI generation requires a trained character model; this Skill automates the Higgsfield Soul training workflow from photo collection to a reusable reference ID. ## Core Features & Use Cases - Soul Training: Submits 5-20 face photos to the Higgsfield CLI and trains a personalized identity model, returning a reusable reference ID. - Variant Selection: Chooses between --soul-2 for image generation and --soul-cinematic for cinematic or video work based on downstream use. - Bootstrap & Error Handling: Verifies CLI installation, authentication, and paid plan status, with guidance for training failures and session expiry. - Use Case: A creator wants their own face in generated videos. They provide a name and 10 varied photos; the Skill trains the Soul and hands back an ID usable via --soul-id in higgsfield-generate. ## Quick Start Ask the assistant to train a Higgsfield Soul named after you using a set of your face photos, for example: "Create my Soul named alex from these photos."

Frequently Asked Questions about higgsfield-soul-id

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

FAQPage Schema
How do I train a Higgsfield Soul character from my photos?

Provide a one-word name and 5-20 face photos, then run higgsfield soul-id create with the --soul-2 or --soul-cinematic flag. The CLI uploads local paths automatically and returns a reference ID once training finishes.

How many photos do I need for Soul training?

Soul training requires a minimum of 5 and a maximum of 20 photos, with 8-12 being the recommended range. Photos should show varied angles, lighting, and expressions with a clear, unobstructed face.

What is the difference between soul-2 and soul-cinematic variants?

The --soul-2 variant is the default and targets image generation, while --soul-cinematic is intended for cinematic and video work. Choose based on your stated downstream use.

Does Higgsfield Soul training work on a free plan?

No, Soul training requires a paid Basic plan or higher. If higgsfield account status shows a free plan, the CLI returns a Minimum Basic plan required error and you must upgrade first.

Why did my Soul training fail?

Training commonly fails due to too few or too uniform photos, heavy occlusion like sunglasses or hats, group photos, or non-image uploads. Swap in better varied photos and retrain.

When should I not use Soul training for face generation?

Do not use Soul training for one-shot face swaps or named non-photo characters. Use higgsfield-generate with --image for single face swaps or a text prompt for fictional characters instead.