Drawing（Doubao image generation）

POST

/doubao/drawing

Image generation model from Doubao
Support model：
general_v2.1_L
This upgrade (compared to General 2.0 and General 2.0 PRO) not only significantly improves structural accuracy and aesthetics but also focuses on enhanced support for Chinese and English character rendering. It is recommended for use in general categories (not limited to portraits or human figures).

general_v2.0_L
"The General 2.0 Pro model (previously known as the General 2.0L model) is a native bilingual model supporting both Chinese and English. It offers further improvements in image-text alignment and aesthetics, making it highly recommended for use in general categories (not limited to portraits or human figures).
Compared to the General 2.0 model (previously known as the General 2.0S model), it delivers enhanced performance with a slight increase in latency."

general_v2.0
The General 2.0 model (previously known as the General 2.0S model) is a native bilingual model supporting both Chinese and English. It features significant improvements in image-text alignment and aesthetics.
It is recommended for use in general categories (not limited to portraits or human figures) and scenarios that demand extremely high-quality results with strict latency requirements.

Introduction：https://www.volcengine.com/docs/6791/1366783

Price：0.05 PTC/call

Request

Authorization

Header Params

Authorization

string

required

Example:

Bearer {{YOUR_API_KEY}}

Body Params application/json

prompt

string

required

"Prompts for generating images can be entered in either Chinese or English.

To ensure higher accuracy for the text you want to display in the image, enclose the desired text in quotation marks within the prompt. For example: 'A Christmas poster with the text "Merry Christmas" on it.'"

model_version

enum<string>

optional

The model version name is fixed as: general_v2.1_L.

Allowed values:

general_v2.1_Lgeneral_v2.0_Lgeneral_v2.0

req_schedule_conf

enum<string>

optional

Default version: general_v20_9B_pe
Standard version: general_v20_9B_rephraser
Aesthetic version: general_v20_9B_pe

Allowed values:

general_v20_9B_pegeneral_v20_9B_rephraser

llm_seed

string

optional

seed

integer

optional

The random seed serves as the basis for determining the initial state of diffusion. The default value is -1 (random).

If the random seed is set to the same positive integer and all other parameters remain identical, the generated images are highly likely to be consistent in appearance.
Default value: -1.

scale

number

optional

The degree to which the text description influences the output is adjustable
Default value：3.5
Value range：[1, 10]

ddim_steps

integer

optional

The number of steps for generating images .Using a higher number of steps may improve image quality but can significantly increase latency, potentially causing service timeouts. It is recommended to stick to the default value for optimal performance.
Default value：25
Value range：[1, 200]
Recommended value range：[1, 50]

width

integer

optional

Width of Generated Image
Default Value：512
Value Range：[256, 768]

height

integer

optional

Height of Generated Image
Default Value：512
Value Range：[256, 768]

use_pre_llm

boolean

optional

Enabling text rephrasing allows the input prompt to be expanded and optimized.
If the input prompt is relatively short, it is recommended to enable this feature.
If the input prompt is longer, it is recommended to disable it.
Default value: true

use_sr

boolean

optional

True: Text-to-image generation + AIGC super-resolution.
False: Text-to-image generation only
Default value: true

sr_seed

number

optional

sr_strength

number

optional

sr_scale

number

optional

sr_steps

number

optional

is_only_sr

boolean

optional

return_url

boolean

optional

Specifies whether the output should include an image link (the link will be valid for 24 hours)

logo_info

object

optional

Watermark information

add_logo

boolean

optional

position

integer

optional

水印的位置，取值如下：
0-右下角
1-左下角
2-左上角
3-右上角
默认0

language

integer

optional

水印的语言，取值如下：
0-中文（AI生成）
1-英文（Generated by AI）
默认0

opacity

number

optional

水印的不透明度，取值范围0-1，1表示完全不透明，默认0.3

logo_text_content

string

optional

明水印自定义内容

Example

{
  "prompt": "千军万马",
  "model_version": "general_v2.1_L",
  "req_schedule_conf": "general_v20_9B_pe",
  "llm_seed": -1,
  "seed": -1,
  "scale": 3.5,
  "ddim_steps": 25,
  "width": 512,
  "height": 512,
  "use_pre_llm": true,
  "use_sr": true,
  "sr_seed": -1,
  "sr_strength": 0.4,
  "sr_scale": 3.5,
  "sr_steps": 20,
  "is_only_sr": false,
  "return_url": true,
  "logo_info": {
    "add_logo": false,
    "position": 0,
    "language": 0,
    "opacity": 0.3,
    "logo_text_content": "这里是明水印内容"
  }
}

Request samples

Shell

JavaScript

Java

Swift

PHP

Python

HTTP

Objective-C

Ruby

OCaml

Dart

curl --location --request POST 'https://api.302.ai/doubao/drawing' \
--header 'Authorization: Bearer sk-jls4AaVBGoe1GwZD64qZA1qyKTN1MPHa4NmvH1cT68z7K1Zz' \
--header 'Content-Type: application/json' \
--data-raw '{
    "prompt": "千军万马",
    "model_version": "general_v2.1_L",
    "req_schedule_conf": "general_v20_9B_pe",
    "llm_seed": -1,
    "seed": -1,
    "scale": 3.5,
    "ddim_steps": 25,
    "width": 512,
    "height": 512,
    "use_pre_llm": true,
    "use_sr": true,
    "sr_seed": -1,
    "sr_strength": 0.4,
    "sr_scale": 3.5,
    "sr_steps": 20,
    "is_only_sr": false,
    "return_url": true,
    "logo_info": {
        "add_logo": false,
        "position": 0,
        "language": 0,
        "opacity": 0.3,
        "logo_text_content": "这里是明水印内容"
    }
}'

Responses

🟢200OK

application/json

Body

completed_at

string

required

created_at

string

required

error

string

required

string

required

model

string

required

output

string

required

started_at

string

required

status

string

required

Example

{
  "completed_at": "",
  "created_at": "",
  "error": "",
  "id": "",
  "model": "",
  "output": "",
  "started_at": "",
  "status": ""
}

Modified at 2024-12-20 02:47:02

Luma-Photon-Flash（Fast image generation）

Imagen-3 (Image generated)