POST /pub/v1/generations
to make the character settings more easier and flexible.mode
is an enum field and can take values:flexible
- Ideal for stylized characters and image generations.balanced
- Ideal for creating realistic images and more.strong
- Character similarity is most important.weight
and enable_facelock
are deprecated and will be removed in the future.mode
to the request body of character in POST /pub/v1/generations
.
media_id
for video generation, eliminating the need to re-upload images as assets.3:5
or 5:3
video_anyone
generation support in the request body of POST /pub/v1/generations
.GET /pub/v1/controlnets?type=flux
and you will see new controlnets for the Flux model.control_net
in the request body of POST /pub/v1/generations
with flux models and styles to generate images with ControlNet/Pose Lockcontrol_net
with Flux Model or Style, please note that the output image’s aspect ratio will automatically match the aspect ratio of the ControlNet reference image (specified by asset_id), regardless of any other aspect ratio settings.resize_mode
and control_mode
are not supported for control_net
when used with Flux Model or Style.flux
model support in the request body of POST /pub/v1/generations
.
flux
style in the request body of POST /pub/v1/generations
.
style_detail
field along with style
in response body of POST /pub/v1/generations
.
type
support in the query param of GET /pub/v1/models
, GET /pub/v1/styles
, GET /pub/v1/controlnets
and GET /pub/v1/loras
.
narrator
to the request body of POST /pub/v1/generations
.
POST /pub/v1/generations
images
renamed to media
image_size
renamed to dim
under data→mediatype
added under data→mediaGET /pub/v1/generations
images
renamed to media
under data in response.size
renamed to dim
under data→media in response.type
added under data→media in responseimage_id
changed to media_id
in Query Parameter