Face Generation
Generate photorealistic faces in seconds and send the best shots straight into other tools.
Written By Arina
Last updated 26 minutes ago
What the Face Generation tool does
Face Generation converts a set of appearance parameters — gender, age, ethnicity, body type, eye color, hair details — into unique portraits in ≈15s. Every result is royalty-free for commercial use — send the best shots straight into Image Upscaler, Head & Face Swap, Image to Video, or Variations workflows.
🎬 Results
A look at what Face Generation can produce.

Your step-by-step guide
Step 1. Surprise Me (optional) — quickly fills the generator with a random combination of settings. Perfect for inspiration, testing different looks, or rapid batches.
Step 2. Gender (required) — choose between Female and Male. Some appearance settings change depending on the selected gender.
Step 3. Age (required) — adjust the slider to define the approximate age of the generated character.
Step 4. Ethnicity (required) — select the character's ethnicity: European, Asian, Latin American, African, or Middle Eastern.
Step 5. Body Type — choose the body shape: Slim, Athletic, Average, Curvy, Muscular, or Plus Size. The result is a face portrait — Body Type influences facial structure and proportions, not the body itself.
Step 6. Eye Color — select the eye color for the generated character. Both natural and stylized shades are available.
Step 7. Hairstyle — choose the hairstyle: Straight, Wavy, Curly, Kinky/Coily, Braided, Bun, or Ponytail.
Step 8. Hair Color — select a natural or fantasy hair color: Light Blonde, Blonde, Honey Blonde, Ginger, Light Brown, Brown, Black, Gray, Green, Blue, Violet, or Pink.
Step 9. Resolution — choose the output resolution (e.g. 2K).
Step 10. Number of Images — select how many faces to generate in this run.
Step 11. Start Generation — starts the generation process using your selected settings. The credit cost per image is shown below the button before you confirm. Fields marked with a red asterisk are required.
Quick-start walkthrough

Choose traits or smash Surprise Me.
Confirm the cost and hit Start Generation. A task tile appears on the left ("facegen-YYYY-MM-DD-HH-MM-ID").
Watch status – Created → Processing → Completed / Failed. Time-outs auto-retry once; persistent fails are flagged.
Click the task tile (or View X/4 Generations) to open the Results page.
Actions on results

Select / Deselect: use the checkbox in each thumbnail corner.
Retry Failed: visible if any of the four render errors out.
Send Selected to Tool or to Publishing: open the mini-launcher for Head & Face Swap, Image Upscaler, Reference-To-Image, Image to Video, or Variations.
Download Selected / Download All: one-click ZIP export.
Create Cloud Archive: store images straight to your connected S3 or Google Drive bucket.
Pro tips
Batch like a pro → queue 50+ jobs; GPU auto-scaling keeps average wait ≈15–25s.
Use Surprise Me three times, favourite the best eight faces, then pipe to Image Upscaler → Variations for a polished "Meet the Team" social post in under five minutes.
Troubleshooting
You're working with AI — occasional mistakes or artifacts are normal, and a 100% correct result isn't guaranteed.
Task failed? Try restarting it using Retry. If the issue persists, contact support and we'll take a look.
Blurry outputs? Send directly to Image Upscaler.
Still stuck? Ping us via the chat bubble or email [email protected].
FAQ
How many faces per run?
How many faces per run?
From 1 to 4.
Do I own the images?
Do I own the images?
Yes — full commercial license, no attribution needed.
Can I upload a reference face?
Can I upload a reference face?
For matching style or expression, yes — use Reference-To-Image after your first run.
What's next
Head & Face Swap — apply this face to other images
Image Upscaler — sharpen and enlarge your result
Reference-To-Image — build variations from this face
Was this helpful?
More in Tools
LoRAs AI EffectsCompanionImage to VideoText to VideoStill need help? Ask the team