Skip to main content
POST
/
pub
/
v1
/
generations
Generate media
curl --request POST \
  --url https://api.rendernet.ai/pub/v1/generations \
  --header 'Content-Type: application/json' \
  --header 'X-API-KEY: <x-api-key>' \
  --data '[
  {
    "aspect_ratio": "1:1",
    "batch_size": 1,
    "cfg_scale": 7,
    "character": {
      "character_id": "chr_01",
      "weight": 0.6,
      "enable_facelock": true,
      "mode": "balanced"
    },
    "control_net": {
      "asset_id": "ast_01",
      "control_mode": 0,
      "name": "Normal",
      "resize_mode": 0
    },
    "facelock": {
      "asset_id": "ast_01"
    },
    "loras": [
      {
        "name": "lora_name",
        "weight": 0.5
      }
    ],
    "model": "JuggernautXL",
    "narrator": {
      "image_asset_id": "ast_imgxxxxx",
      "video_asset_id": "ast_vidxxxxx",
      "audio_asset_id": "ast_audxxxxx",
      "script": "Thank you for trying Narrator",
      "voice": "Rachel"
    },
    "prompt": {
      "negative": "nsfw, deformed, extra limbs, bad anatomy, deformed pupils, text, worst quality, jpeg artifacts, ugly, duplicate, morbid, mutilated",
      "positive": "a man looking into the camera"
    },
    "quality": "Plus",
    "sampler": "DPM++ 2M Karras",
    "seed": 1234,
    "segment": {
      "asset_id": "ast_01",
      "find": "red shirt",
      "replace": "blue shirt"
    },
    "steps": 20,
    "style": "Bokeh",
    "style_detail": {
      "name": "Black & White",
      "base_model": "flux"
    },
    "true_touch": {
      "asset_id": "ast_01",
      "enhance_strength": 0.5,
      "scale_factor": 1.5
    },
    "video_anyone": {
      "media_id": "ast_01",
      "prompt": "Camera static, a woman riding a bicycle",
      "duration": 5
    }
  }
]'
{
  "data": {
    "credits_remaining": 20,
    "generation_id": "gen_01",
    "media": [
      {
        "id": "model_01",
        "dim": {
          "height": 512,
          "width": 512
        },
        "model": "JuggernautXL",
        "status": "initiated",
        "style": "Bokeh",
        "style_detail": {
          "name": "Black & White",
          "base_model": "flux"
        },
        "type": "image"
      }
    ],
    "result": "initiated"
  },
  "err": {}
}
Start using mode in character object while generating images using character.
weight and enable_facelock are deprecated and will be removed in the future.

Example:
[
    {
        .
        .
        "character": {
            "character_id": "char_001",
            "mode": "balanced"
        }
        .
        .
    }
]
Added support for Video Anyone generation
Any attempt to produce NSFW, violent, or public figure content will result in permanent account termination.
You can use either a generated image’s media_id or an uploaded image asset’s asset_id as the input image identifier media_id for video generation, eliminating the need to re-upload images as assets.
Example request
[
    {
        "video_anyone": {
            "media_id": "ast_xxxxxxxx",
            "prompt": "A person riding a bicycle",
            "duration": 5
        }
    }
]
For video_anyone generation make sure that input image meets the following conditions to avoid generation failure:
  1. Input image aspect ratio should be either 3:5 or 5:3.
  2. Maximum characters permitted for prompt is 500 only.

When using the generation array, ensure that each object contains only one of the following properties: segment, style, model, true_touch, narrator or video_anyone. These properties have a priority order as follows:
  1. segment
  2. style
  3. model
  4. true_touch
  5. narrator
  6. video_anyone
If multiple properties are provided, the one with the highest priority will be used.
The narrator can be generated either by using a script paired with a voice(List all the available voices using GET /pub/v1/voices endpoint) or by utilizing an existing audio asset.
segment, true_touch, narrator and video_anyone generations only acknowledge parameters that are passed within their own objects.
For narrator generation make sure that video asset meets the following conditions to avoid generation failure:
  1. The maximum supported frame rate for a video asset is 30 FPS, calculated by dividing the total number of frames by the total duration.
  2. Maximum height and width of the video asset should be 1080 and 1920 respectively.

Headers

​
X-API-KEY
string
required

API key needed to access our public endpoints. You can find yours under the 'Account' section of the website

Body

application/json · object[]
​
aspect_ratio
enum<string>

The aspect ratio of the image for the generation

Available options:
1:1,
2:3,
3:2,
4:5,
16:9,
9:16
Example:

"1:1"

​
batch_size
integer

The batch size of the generation (Max permissible value depends on your subscription plan). Visit RenderNet pricing for more details

Example:

1

​
cfg_scale
number

AI guidance for this generation. Higher the value, the output will be closer to the prompt (but may result in distorted images after a point)

Required range: 4 <= x <= 12
Example:

7

​
character
object

If you want to create images of a character, use the character’s name in the positive prompt. Example: {character_name} riding a bike

​
control_net
object

The control net(pose control) for the generation if you want to use (optional)

​
facelock
object

The face you want to restore for the generation if you want to use (optional)

​
loras
object[]

Query the list resources endpoint GET /pub/v1/loras to view all available LoRAs. You can add multiple LoRAs to a single generation. Make sure your LoRAs have the same base model (SD 1.5 / SDXL) as your selected style/model.

​
model
string

Query the list models endpoint GET /pub/v1/models to view all available models.

Example:

"JuggernautXL"

​
narrator
object

For Narrator generation, pass either script or audio_asset_id along with other required details in narrator payload.

​
prompt
object

The prompt for the generation

​
quality
enum<string>

The quality of the image for the generation (Case Sensitive)

Available options:
Plus,
Regular
Example:

"Plus"

​
sampler
enum<string>

The sampler you want to use for the generation (Case Sensitive)

Available options:
DPM++ 2M Karras,
DPM++ 2M SDE Karras,
DPM++ 2S a Karras,
DPM++ SDE,
DPM++ SDE Karras,
Euler a
Example:

"DPM++ 2M Karras"

​
seed
integer

The seed for the generation (randomized if you don't want to use a seed)

Example:

1234

​
segment
object

Changes the input asset image based on the find and replace prompts

​
steps
integer

The number of steps you want AI to take for the generation

Required range: 10 <= x <= 30
Example:

20

​
style
string

Query the list styles endpoint GET /pub/v1/styles to view all available styles.

Example:

"Bokeh"

​
style_detail
object

Query the list styles endpoint GET /pub/v1/styles to view all available styles.

​
true_touch
object

Enhances and upscales the input image passed in the asset_id

​
video_anyone
object

Convert image to video seamlessly.

Response

201 - application/json

A list of submitted generation requests

​
data
object
​
err
object

The error if the request was not successful