What this flow does
This is a do-it-all finisher for reference-driven shoots. For each uploaded reference image, we:
Analyze the photo (pose, framing, lighting, styling) and generate similar—but not identical—shots.
Upscale the results for detail and clarity (you choose Basic for speed or Premium for higher realism).
Face Swap every image to the same source face so the identity is consistent across the entire batch.
Use it when you know the look (reference) and the face (Persona) you want—and you need finished images fast.
When to use this flow
You found a great reference (Pinterest/IG/your own photo) and want a near-match pose/location/mood with variations.
You’re producing a campaign batch and want identity locked across 10–100 images.
You prefer one click instead of hopping between tools (Gen by Ref → Upscale → Face Swap).
If you still plan heavy edits (dress-up, background replace), do those first and run this flow last.
Inputs (what each control means)
Upload a Face Image
Upload a clear, high-quality face image that will be used as the source identity for face swap.
Supports a single image or selection from Personas. Use a sharp front or slight 3/4 portrait with minimal occlusion.Upload Ref Images
Upload reference images to be processed with upscaling and face swap.
Supports up to 100 files per batch. Each reference image generates one final output.Reference Strength
Controls how closely the generation follows the reference image.
Lower values allow more creative variation; higher values enforce stricter pose and composition matching.Positive Prompt (optional)
Describe what you want to see in the final image, such as lighting, mood, background, or visual details.
Example: “soft window light, shallow depth of field, clean background”.Negative Prompt (optional)
Exclude unwanted artefacts or elements such as: “flicker, plastic skin, heavy sharpening, watermark, text”.LoRA (optional)
Select an optional LoRA to influence style or appearance.
Leave set to None if no additional stylistic control is needed.Upscaler Version
Choose the upscaling quality preset:
Basic = faster/cheaper; Premium = more realism/micro-detail (slower/costlier). Face Swap quality is identical in both; this choice only affects the upscale.Start Generation
Starts the flow generation process.
Credit cost is displayed below the button (3 credits per image).
Quick start (3 minutes)
Upload Ref Images (up to 100 files).
Set Reference Strength around the middle; add a concise Positive Prompt and a tight Negative Prompt (optional).
(Optional) Select a LoRA if you need additional style control.
Upload a Face Image or select a Persona to apply the identity.
Choose Upscaler: Basic (speed) or Premium (finals).
Click Start Generation — each reference produces four final images, already upscaled and face-swapped.
Pro tips for high-quality, consistent results
Curate references – use clean edges, clear poses, and simple backgrounds. One reference per pose works better than collages.
Avoid occlusions – hair covering the face, hands near eyes, or heavy accessories can reduce face swap accuracy.
Batch smart – group references by scene (beach set, office set). Consistent lighting improves identity matching.
LoRAs – mix at most 1–2; start with moderate strength and increase only if necessary.
Premium for finals – use Basic to explore, Standard Enhance / Premium to finish your keepers.
Naming – include persona, scene, and output size in job names for easier tracking across flows.
Face swap stability – for the most reliable identity preservation, choose Basic or Premium Safe Face upscaling modes.
Common issues & fixes
Identity looks off/uncanny → Use a sharper, front-lit face source; avoid sunglasses; raise reference Strength a bit so pose/lighting stabilize.
Skin tone mismatch → Prefer Premium upscale (better tone harmonizing) and choose references with similar lighting to your face source.
Artefacts with multiple LoRAs → Reduce LoRA strengths or use only one; higher strengths increase artefact risk.
Too close to the reference (not enough variation) → Lower Strength slightly and add a few prompt hints (e.g., “wider lens feel”, “different hand placement”).
FAQ
How many outputs do I get?
Each reference image produces four final outputs.
Can I mix different faces in one run?
No. Each task supports one source face only for maximum identity consistency. Run separate tasks for different Personas or face images.
Do LoRAs work here?
Yes (SDXL LoRAs). Use them to guide physique or overall style. LoRA tokens are injected automatically — keep your prompts clean and human-readable.
What does the Upscaler choice change?
Only the enhancement stage.
Basic is faster and more economical.
Premium adds more realism and micro-detail.
Why creators love this flow
It replaces a 3-step pipeline with one job, producing cohesive, high-detail images where the same face appears across every shot. Perfect for campaigns, lookbooks, and weekly content packs.