How to Create SFW Content for a New AI or Real Model

Learn how to build a consistent, studio-quality photo set for a brand-new AI or real model. This step-by-step guide walks you through Text-to-Image, Generation-by-Reference, and Image Editor workflows.

ZenCreator Team
Updated November 21, 2024
5 min read
workflowtext-to-imageface-generationsfw

Creating a new model means producing a consistent photo set — 10 to 20 cohesive images of the same person or character in different poses, lighting, and angles.

You'll later use this set for LoRA training (soon), PhotoShoots, or future Gen-by-Reference sessions.

🧭 Available Workflows

You can build your first SFW photo set in three main ways:

  1. Text-to-ImageFace GenFace Swap
  2. Image-to-ImageFace GenFace Swap
  3. Image EditorFace Gen

1️⃣ Text-to-Image → Face Gen → Face Swap

Goal

Generate clean, realistic photos from text prompts, upscale them, then replace faces with your generated model's face for perfect consistency.

Step 1 · Generate Base Images

Tool: Generator by Prompt Recommended modes:

  • WAN (SFW) – realistic, balanced lighting
  • SDXL – stylised / body-shaping freedom
  • (Avoid General NSFW for SFW work)

Try to create:

  • 3 full-body shots (from different angles)
  • 3 portraits
  • 3 sitting poses
  • 3 lying poses All with varied lighting and locations.

🎯 Tips

  • Describe appearance in detail (hair color, age, body type, mood).
  • Mention lighting and emotion for more life-like results.
  • To expand angles later, you can send images to the Carousel Generator.

📸 Example Prompt full-body photo of a 20-year-old American woman with long white hair, standing barefoot in a cozy kitchen, morning light through the window, holding a mug, cinematic tone, DSLR realism

Step 2 · Select and Upscale

Tool: AI Upscaler Use Premium or Face-Safe mode for clean detail recovery. Choose the most consistent shots (10–20 total). No need to preserve the original face —we'll swap it next.

Step 3 · Generate the Face

Tool: AI Face Generator Define gender, age, ethnicity, and style to match your prompt. Generate 4 faces per run; star your favorite one.

Step 4 · Apply Face Swap

Tool: AI Face Swap Upload the upscaled images as targets and the new face as reference. Run bulk swap for a consistent look across all shots. Result: a 10–20 photo SFW set with identical identity.

2️⃣ Image-to-Image → Face Gen → Face Swap

Goal

Use reference photos (real or AI) to re-imagine them as your model's own shoot.

Step 1 · Pick References

Choose a real person or photos from Instagram that match the look you want. Keep hair and complexion close to your target model.

Step 2 · Generate by Reference

Tool: Generator by Reference Model options: Flux or SDXL (for SFW). Optionally attach LoRas to adjust body shape. Generate variants and select the most consistent results.

Step 3 · (Optionally) Upscale and Face Swap

Use the same Upscale and Face Swap steps from Workflow #1 to finalize the set.

3️⃣ Image Editor → Face Gen

Goal

Combine a face, body, and environment into a controlled, composed shot.

Step 1 · Prepare Source Images

  • Model face
  • Body photo (clear pose and silhouette)
  • Optional references for clothes or background
  • A precise prompt describing scene and mood 💡 Use simple backgrounds — the editor copies visible details from references.

Prepare source images

Step 2 · Launch Image Editor

Feed the prepared inputs (face, body, environment). Generate until you reach your desired composition and lighting.

Step 3 · Refine and Expand

Select the best shots, then optionally send them to the Carousel Generator to create angle variations.

📦 Next Steps

Once you have your SFW photo set:

  • Train a LoRA for advanced personalization (soon).
  • Generate new photos by reference or prompt.
  • Schedule content via the Publishing Hub.