Generate Images

Action GenerateImages

Introduction

This interface is used to generates images.

Request

Parameter Name	Type	Description	Required
Action	`String`	The corresponding API instruction name	✓
PublicKey	`String`	User's public key, which can be obtained from the PICPIK User Center (opens in a new tab)	✓
Signature	`String`	User's signature generated based on the public key and API instruction, see Signature for details	✓
request_uuid	`String`	Request UUID, randomly generated by using the UUID generation method. This parameter can be utilized to manage the requests of your business and clients.

Parameter Name	Type	Description	Required
MainModel	`Model`	Main model for generating images	✓
SubModelList	`[]Model`	The submodel list for generating images includes LoRA, Embedding, and HyperNetwork.
PositivePrompts	`[]String`	Content description of what you want to appear in the image.
NegativePrompts	`[]String`	Content description of what you do not want to appear in the image.
TranslatePrompts	`Bool`	Translate other languages to English, see support language list here. Default value: true
Seed	`Int`	The number of a seed (-1~2147483647), -1 represents a ramdom seed. Default value: -1
Width	`Int`	Image width (64~2560). Default value: 1024
Height	`Int`	Image height (64~2560). Default value: 1024
BatchSize	`Int`	Batch size for generating images (1~16). Default value: 1
UseMainModelVAE	`Bool`	Whether to use the codec that comes with the main model. Default value: false

Parameter Name	Type	Description
ControlNetList	`[]ControlNet`	The ControlNet models (up to three) and parameters list.
Image2Image	`Image2Image`	Parameters related to a reference image when generating from it.
Tileable	`Bool`	Wether to generate tileable images or not. It is mostly used to create textures or patterns that can be tiled together. Default value: false
ImageAndMaskList	`[]Base64`	A list of reference images and masks encoded in Base64.

Intro to Complex Data Types

`Model`

Field Name	Type	Description	Required
Name	`String`	Model name
ModelCode	`String`	Model code	✓
IsOfficial	`Bool`	Is official model or not
TriggerWordList	`[]String`	Trigger word list, only available with LoRA model
InferenceParams	`InferenceParams`	Inference parameters

`ControlNet`

Parameter Name	Type	Description	Required
ImageIndex	`Int`	The index of an image in ImageAndMaskList, starting from 0.	✓
MaskIndex	`Int`	The index of a mask image in ImageAndMaskList, starting from 0. A mask can be drawn independently or obtained through the Remove-Background or Content Extraction APIs.
InvertMask	`Bool`	Wether to invert the mask. Default: false
PreProcessor	`PreProcessor`	The model and configuration information of a pre-rocessor.	✓
ControlModel	`ControlModel`	The model and configuration information of the ControlNet.	✓

`Image2Image`

Parameter Name	Type	Description	Required
ImageIndex	`Int`	The index of an image in `ImageAndMaskList` or `ImageList`, starting from 0.	✓
ReferStrength	`Float`	The reference strength of this image (0~1). Default value: 0.5

`PreProcessor`

Parameter Name	Type	Description	Required
ModelCode	`String`	The model code of a ControlNet pre-processor.	✓
InferenceParams	`InferenceParams`	Inference parameters of a pre-processor

`ControlModel`

Parameter Name	Type	Description	Required
ModelCode	`String`	ControlNet model code	✓
InferenceParams	`InferenceParams`	Inference parameters of a ControlNet model

`InferenceParams`

Note: Specific parameters are only valid for a particular model data type (parent) and are not generic.

Parameter Name	Type	Valid Parent Type	Description	Default	Range	Step
LoraWeight	`Float`	`Model`	The influence weight when inferencing. Only available with the LoRA model.	1	-1～2	0.01
Resolution	`Int`	`PreProcessor`	The resolution of images generated by the pre-processor.	512	64~2048	64
ControlWeight	`Float`	`ControlModel`	The influence weight of a ControlNet model when inferencing.	1	0~2	0.01
StartingStep	`Float`	`ControlModel`	Timing (ratio) of a ControlNet model to start intervening in the inference process.	0	0~1	0.01
EndingStep	`Float`	`ControlModel`	Timing (ratio) of a ControlNet model to stop intervening in the inference process.	1	0~1	0.01

Parameters-Passing Tip

💡

A Model in the MainModels list retuned by the ListModels interface can be assigned to the MainModel parameter of this interface directly, as demonstrated in Python:

# Assume a variable "sdModels" as the JSON data returned by the interface "ListModels"
{
    "MainModel": sdModels["MainModels"]["PICPIK General"],
}

Parameters that require model data, such as SubModelList and ControlNetList, can also be set by utilizing this tip.

Simple Request Example

{
	"Action": "GenerateImages",
	"PublicKey": "ALLaXAM4cNdHXPDJAJJ4k8Hh1hZS2EgJF1D0FlPwfl",
	"Signature": "83f3ce02e110bbb25cecb5092637934a95e3bdc8",
	"request_uuid": "432c57d3-102b-4817-9e98-28054b33db8e",
	"MainModel": {
		"ModelCode": "PICPIK Anime"
	},
	"PositivePrompts": [
		"a cat on the sofa"
	]
}

Advanced Request Example

{
    "Action": "GenerateImages",
    "PublicKey": "ALLak9M4cNdHXKOJAJJ4k8Hh1hZS2EgJF1D0FlPwfl",
    "Signature": "59734dfd3985e96d452a2072533f47c766d52c19",
    "request_uuid": "9e3f5da5-8020-463c-af54-a24912f4eb0e",
    "MainModel": {
        "ModelCode": "PICPIK General"
    },
    "SubModelList": [
        {
            "InferenceParams": {
                "LoraWeight": 1
            },
            "ModelCode": "lora_model1_20230825_165436"
        }
    ],
    "NegativePrompts": [],
    "PositivePrompts": [
        "一只猫在沙发上"
    ],
    "TranslatePrompts": true,
    "Width": 512,
    "Height": 512,
    "BatchSize": 1,
    "Seed": -1,
    "Image2Image": {
        "ImageIndex": 0,
        "ReferStrength": 0.5
    },
    "ControlNetList": [
        {
            "ImageIndex": 1,
            "MaskIndex": 2,
            "InvertMask": false,
            "PreProcessor": {
                "InferenceParams": {
                    "Resolution": 512
                },
                "ModelCode": "mlsd"
            },
            "ControlModel": {
                "InferenceParams": {
                    "ControlWeight": 1,
                    "EndingStep": 1,
                    "StartingStep": 0
                },
                "ModelCode": "control_v11p_sd15_mlsd"
            }
        }
    ],
    "Tileable": false,
    "UseMainModelVAE": false,
    "ImageAndMaskList": ["data:image/png;base64,xxx","data:image/png;base64,xxx","data:image/png;base64,xxx"]
}

Response

Field Name	Type	Description	Required
RetCode	`Int`	Return status code: 0 indicates a successful return, while non-zero indicates a failure	✓
Action	`String`	Instruction name	✓
Message	`String`	The error message is returned when `RetCode` is non-zero, providing a detailed description
request_uuid	`String`	Returns the UUID that was included in the request

Parameter Name	Type	Description	Required
TaskUID	`String`	Task UID, used to query task details and get results	✓

Response Example

{
	"Action": "GenerateImages",
	"RetCode": 0,
	"Message": "",
	"request_uuid": "432c57d3-102b-4817-9e98-28054b33db8e",
	"TaskUID": "ckt3uqlrh4miq8mr9lgg"
}

List ControlNet Models Generate Images (SDXL)