VirtuaVixen's Scene Workflows represent the most advanced category of AI porn generation available today. While standard I2V (Image-to-Video) workflows produce a single clip from a single image, Scene Workflows generate 4 connected action segments from one input image — producing a complete ~19-second scene with narrative progression. If you're new to AI-generated adult content, start with our complete guide to making AI porn for the fundamentals.

This article breaks down what Scene Workflows are, how they work under the hood, and what each of the 16 available scenes generates.

The Problem Standard Workflows Solve (And Where They Fall Short)

Standard I2V workflows generate one action from one image. A blowjob workflow makes a blowjob clip. A cowgirl workflow makes a cowgirl clip. If you want a scene with progression — oral leading into sex leading into a finish — you need to generate each clip separately, then stitch them together manually in an editor.

That approach has problems. Each clip starts from a different generation seed, so character appearance can drift between segments. Transitions feel disjointed. You spend more time editing than creating. Scene Workflows solve all of this by generating all 4 segments from a single image in one pass, maintaining character consistency throughout the entire sequence.

The Technical Breakdown

Each Scene Workflow runs on the following stack:

• Base checkpoint: DasiwaWAN22I2V14BLightspeed with SynthSeduction merge — a speed-optimized WAN 2.2 14B model tuned for NSFW content
• Video node: WanImageToVideoSVIPro — generates 4 segments of 81 frames each from the input image
• Base LoRA stack: SVI_v2_PRO + 4llinon — applied to every scene for motion quality and anatomical consistency
• Action-specific LoRAs: Each scene loads additional LoRAs tuned for its particular action type (oral, anal, riding, etc.)
• 4 editable prompt fields: Each segment has its own prompt, so you can customize exactly what happens in each stage of the scene
• Sampler: LCM sampler, beta scheduler, 6 steps, CFG 1 — fast inference without quality loss
• Total output: ~19 seconds of continuous video with built-in scene progression

The key innovation is the WanImageToVideoSVIPro node. Instead of running 4 separate generation passes, it produces all segments in a single forward pass while sharing the conditioning from your input image. This keeps the character's face, body, and lighting consistent across all 4 segments — something you can't achieve by generating clips individually.

Workflow Specs

• Input: Single image (portrait or full body, 768×1024+ recommended)
• Output: 4 video segments (~4.75 seconds each, ~19 seconds total)
• Frames per segment: 81
• Architecture: WAN 2.2 I2V 14B (Lightspeed variant)
• Sampler: LCM / beta / 6 steps / CFG 1
• LoRA layers: 2 base LoRAs + 1-2 action-specific LoRAs per scene
• Prompt fields: 4 (one per segment, fully editable)
• Special variant: Full Service uses multi-pass PainterI2V architecture for extended scene complexity

Available Scene Workflows

There are 16 Scene Workflows in the AI Porn Studio, each designed for a specific type of scene progression. Here's what each one generates:

1. Oral Scene — Classic oral progression: teasing, licking, sucking, and deepening intensity across 4 segments. The go-to choice for POV blowjob content with realistic head movement and pacing.
2. Titjob Scene — Breast-focused action with natural compression physics. Progresses from teasing through full titjob motion with increasing speed and intensity.
3. Free Prompt Scene — No predefined action. You write all 4 segment prompts from scratch. Maximum creative control for custom scene ideas that don't fit other categories.
4. Cum Shower Scene — Multi-angle cumshot and facial sequence. Builds through anticipation into an extended finish with realistic fluid physics across all segments.
5. Futanari Scene — Futa-specific scene progression. Handles the unique anatomical requirements of futanari content with dedicated LoRAs for consistent rendering.
6. Deepthroat Scene — Aggressive oral with deep insertion focus. Each segment increases depth and intensity, with proper gagging physics and facial expression progression.
7. Anal Scene (Zoom) — Close-up anal with tight camera framing. Optimized for detail-focused content where the camera stays close to the action throughout all 4 segments.
8. Anal Scene — Full-frame anal with wider angles. Captures more body movement and positioning changes as the scene progresses from entry through full action.
9. Cock Worship Scene — Teasing, licking, kissing, and adoration-focused oral content. Slower pacing than the standard Oral Scene, emphasizing eye contact and sensual movement.
10. Sex & Squirt Scene — Penetration building to a squirting climax. The 4-segment structure works particularly well here, with natural build-up leading to an intense finish.
11. Sex & BJ Scene — Mixed scene combining oral and penetration. Transitions between blowjob and sex across segments for a varied, realistic encounter.
12. Missionary Scene — Classic missionary position with natural motion and camera work. Progresses through tempo changes with realistic body mechanics and facial expressions.
13. Cowgirl Scene — Riding position with bounce physics and hip movement. Each segment varies the rhythm and angle for a dynamic, engaging sequence.
14. Sex Scene — General-purpose penetration scene. Versatile enough for various positions and angles — use the 4 prompt fields to define your own progression.
15. BJ & Facial Scene — Oral action building to a facial finish. The segment structure naturally supports the build-up / action / climax / aftermath arc.
16. Full Service — The most advanced scene workflow. Uses a multi-pass PainterI2V architecture instead of the standard single-pass approach. This enables more complex scene transitions and longer narrative arcs across the 4 segments, combining multiple action types in one continuous sequence.

How to Use Scene Workflows

1. Start with a strong input image. Scene Workflows maintain your character across all 4 segments, so the quality of your input directly affects the entire output. Use a clear, well-lit portrait or full-body shot at 768×1024 or higher.
2. Pick the right scene for your content. If you're building a library around a specific character, try multiple scene types with the same input image. One portrait can generate 16 completely different scenes.
3. Edit the segment prompts. Each scene comes with default prompts, but the 4 prompt fields are fully editable. Customize the action, pacing, camera angle, or expressions for each segment to match your creative vision.
4. Use Free Prompt Scene for custom ideas. If none of the 15 predefined scenes match what you want, the Free Prompt Scene gives you a blank canvas with the same 4-segment architecture.
5. Combine clips in post-production. While each scene is designed to work as a standalone ~19-second clip, you can stitch multiple scenes together in CapCut or DaVinci Resolve to create longer videos. A 3-scene edit gives you nearly a minute of content from one input image.
6. Try Full Service for premium content. The multi-pass PainterI2V architecture produces the most complex and varied scene progression. If you're creating content for monetization, Full Service generates the kind of multi-action sequences that perform best on platforms like Fanvue.

Where to Access Scene Workflows

All 16 Scene Workflows are available in the VirtuaVixen AI Porn Studio. Upload an image, select a scene, and generate — no ComfyUI setup, no GPU required.

For creators who prefer to run workflows locally, all scenes are also included in the VirtuaVixen Pro Workflow Pack for ComfyUI.

Browse the full workflow collection: All WAN 2.2 Workflows