GenerateImageReplicateParams

ug-js-sdk

---

ug-js-sdk / GenerateImageReplicateParams

Interface: GenerateImageReplicateParams

Defined in: conversation-manager/types.ts:116

Parameters specific to Replicate image generation. Supports LoRA weights for custom styles.

Examples

const params: GenerateImageReplicateParams = {
  prompt: 'Portrait in watercolor style',
  model: 'flux-schnell',
  aspect_ratio: '3:4',
};

const params: GenerateImageReplicateParams = {
  prompt: 'A landscape',
  lora_scale: 1.0,
};

Properties

prompt

prompt: string

Defined in: conversation-manager/types.ts:118

The text prompt describing the image to generate

---

negative_prompt?

optional negative_prompt: string

Defined in: conversation-manager/types.ts:120

Text describing what to avoid in the generated image

---

aspect_ratio?

optional aspect_ratio: string

Defined in: conversation-manager/types.ts:122

Aspect ratio for the generated image (e.g., '16:9', '1:1')

---

seed?

optional seed: number

Defined in: conversation-manager/types.ts:124

Seed for reproducible generation

---

inference_steps?

optional inference_steps: number

Defined in: conversation-manager/types.ts:126

Number of inference steps for the generation process

---

model?

optional model: string

Defined in: conversation-manager/types.ts:128

Specific model to use for generation

---

image?

optional image: string

Defined in: conversation-manager/types.ts:130

Base64 encoded image for image-to-image generation

---

strength?

optional strength: number

Defined in: conversation-manager/types.ts:132

Strength of the transformation for image-to-image (0.0 to 1.0)

---

guidance_scale?

optional guidance_scale: number

Defined in: conversation-manager/types.ts:134

Guidance scale for generation (higher = more adherence to prompt)

---

lora_weights?

optional lora_weights: string

Defined in: conversation-manager/types.ts:136

URL to LoRA weights

---

lora_scale?

optional lora_scale: number

Defined in: conversation-manager/types.ts:138

Scale factor for LoRA weights (typically -1 to 3)