Reference-To-Image

Upload a reference, set a few options, and generate multiple alternatives you can immediately upscale, face-swap, or publish.

Written By Arina ZenCreator

Last updated 5 days ago

What the tool does

Reference-To-Image analyzes your image (subject, pose, camera angle, lighting, color grade, environment) and produces similar, not identical photos. It keeps the feel of the reference — location, pose family, styling — while exploring natural variations so you can pick the strongest frame.

Great for: recreating a look in different takes, matching a feed's aesthetic, building sets for Variations, and preparing shots for Head & Face Swap to lock identity.

🎬 Results

A look at what Reference-To-Image can produce.

Don't want to set everything up yourself?

Skip the setup — pick a ready-made style from the Templates library and drop in your reference. Templates can be filtered by tags or searched by name.

Your step-by-step guide

Step 1. Model — select the model type for generation. Tap one below to see what it's best at.

Quick pick:

  • Need to combine several people into one scene (Character swap) → Seedream 5 Pro

  • Need to place a face into a reference scene (Scene mode) → Seedream 5 Pro

  • Need the most realistic lighting and texture → Seedream 5 Pro

  • Need better body anatomy, pronounced shapes, or NSFW-focused content → SDXL

  • Need detailed body rendering and anatomically accurate figures → SDXL

Step 2. Mode (Seedream 5 Pro) — Scene: the identity from a face photo is placed into the reference. Character swap: replace the characters in a source scene with your own.

Step 3. Resolution (Seedream 5 Pro) — output resolution tier. 2K produces larger, more detailed images; 1K is faster and cheaper.

Step 4. Aspect Ratio — defaults to Auto (from reference), matching the aspect ratio of your uploaded reference.

In Scene mode:

  • Keep Pose — when on, the reference is sent to Seedream with a fixed prompt: pose, composition, outfit, body, and background all come from the reference, and only the identity changes. Aspect ratio always matches the reference in this case.

  • Reference — upload reference image(s) for pose, composition, or style. Up to 1,000 files.

  • Face — required. The face photo of the character to place into the reference scene. Single file.

  • Body — optional. A body photo used to guide the character's body. Single file.

In Character swap mode:

  • Source Scene — required. Upload the source scene(s) to edit — characters are swapped into these. Up to 1,000 files.

  • Characters — required. Upload 1–4 characters to swap into the source scene.

Step 5. Start Generation — launch the batch. The per-image credit cost appears next to the button.

Working with results

  • Use the checkboxes on thumbnails to Select/Deselect.

  • Download Selected / Download All to export a ZIP.

  • Send to another tool to continue the pipeline (Image Upscaler, Head & Face Swap, Image to Video, Variations).

  • Retry Failed appears automatically if any frames mis-rendered.

Pro tips

Quality in → quality out. Choose references with sharp faces, simple backgrounds, and consistent light.

Batch for choice. Generate 8–10 variations per reference to find a hero frame.

Use LoRAs for physique/style. You don't need a reference with a perfectly consistent body — SDXL + LoRAs will guide it.

FAQ

Troubleshooting

You're working with AI — occasional mistakes or artifacts are normal, and a 100% correct result isn't guaranteed.

Want to make visible edits to the photo (change clothing, background, pose, etc.)? We recommend using Image Editor instead.

Need help choosing the right flow for your shoot? Ping the chat bubble or email [email protected].

What's next