Restore faces
Fix blurry and damaged faces
GFPGAN rebuilds sharp eyes, skin and hair on small or blurry faces, and leaves the rest of the photo as it was.
$0.004
per run in the cloud
$0
on your own machine
A portrait shrunk and blurred, then restored. Strength 0.8.
Every face in the photo
Faces are found and restored one by one, up to ten.
A strength slider
Blend the restored face with the original to keep it natural.
Great after an upscale
Upscale first, then restore the faces.
Reference
Inputs and parameters
The same on every surface: the desktop app, Claude through MCP and the API. Unknown parameters are refused, so a typo never runs with a default.
Inputs
| Name | Type | Limits | Notes |
|---|---|---|---|
| image required | image | 30 MB, 16 MP | Source image |
Parameters
| Name | Type | Default | Accepted | Notes |
|---|---|---|---|---|
| strength | float | 0.8 | 0 to 1 | Blend between the original (0) and the restored face (1). |
| max_faces | int | 10 | 1 to 20 |
Outputs
- imageimage, png
API
Run Restore faces from the API
curl https://imference.com/tools/restore-faces \ -H "Authorization: Bearer $IMFERENCE_API_KEY" \ -F image=@image.png
Files go in multipart fields named after the inputs; JSON works too, with {"inputs": {...}, "params": {...}} and https URLs or base64 data: URIs. The answer is a request_id: poll GET /status?request_id=... until it returns the outputs. Pay per call from a wallet instead at POST /ondemand/tools/restore-faces. A run that fails is refunded to your credits. Details in the reference.
Specs
The details
- Price
- $0.004 · 4 credits
- Runs
- Local and cloud
- Download
- 340.5 MB, once
- License
- Apache-2.0
- Tool id
- restore-faces 1.0.0
- MCP tool
- restore_faces
More tools
Often used with it
Convert / compress
Convert to PNG, JPEG or WebP and set the compression quality.
$0.001
Crop
Crop an image to a rectangle or to an aspect ratio (centered).
$0.001
Describe image
Write a detailed caption of an image, ready to reuse as a prompt (Florence-2).
$0.002
Magic eraser
Remove what the mask covers and fill it with matching background (LaMa).
$0.004
FAQ
Good to know
Does it change the person?
It sharpens the features that are there. Lower the strength to stay closer to the original.
Does it work on old photos?
Yes, on scanned prints too. Scratches outside the faces are not repaired.