API Documentation

Wanx 2.1 - LoRA Features

Generate stylized videos using LoRA models for specific artistic styles and themes

txt2video-14b-lora

$0.28

Generate stylized videos from text descriptions using specific LoRA models for artistic styles.

Text + LoRA Style → Video

img2video-14b-lora

$0.28

Transform images into stylized videos using LoRA models for consistent artistic themes.

Image + LoRA Style → Video

Quick Start

Base URL

https://api.qubico.ai/v1/task

Header Params

X-API-Key: YOUR_API_KEY

Model ID

Qubico/wanx

Parameters Reference

Complete parameter documentation for LoRA-enhanced video generation

Required

model

string

The model identifier for Wanx video generation

"Qubico/wanx"

task_type

string

The LoRA-enhanced task type

"txt2video-14b-lora" | "img2video-14b-lora"

prompt

string

Detailed description of the video content

"A girl with a playful smile, anime style..."

lora_settings

array

LoRA configuration settings

lora_typeLoRA model type (e.g., "ghibli", "anime")

lora_strengthStrength value (0.1-2.0, default: 1.0)

Task-Specific

image

string

img2video-lora only

Input image URL or base64 (required for img2video-lora)

"https://your-image.com/image.jpg"

Optional

negative_prompt

string

Description of what to avoid

"chaotic, distortion, morphing, blurred details"

aspect_ratio

string

Video aspect ratio (default: "16:9")

"16:9" | "9:16"

Available LoRA Types

ghibli

Studio Ghibli animation style

Best for: Whimsical, hand-drawn animation aesthetics

Recommended strength: 0.8-1.2

anime

General anime/manga style

Best for: Character-focused animations with anime aesthetics

Recommended strength: 1.0-1.5

LoRA Strength Guidelines

• 0.1-0.5: Subtle style influence
• 0.6-1.0: Balanced style application
• 1.1-1.5: Strong style emphasis
• 1.6-2.0: Maximum style intensity

Best Practices

• Start with strength 1.0 and adjust as needed
• Use style-specific prompts for better results
• Combine with appropriate negative prompts
• Test different strengths for optimal output

Code Examples

Text-to-Video with LoRA

{
  "model": "Qubico/wanx",
  "task_type": "txt2video-14b-lora",
  "input": {
    "prompt": "A girl with a playful smile leans forward on the ferry railing, holding onto her sunhat as the wind threatens to carry it away. She looks directly at the camera with bright, excited eyes, her laughter almost visible in her expression. The sun reflects off the sparkling water, and behind her, the coastline of a charming island town approaches. The camera slowly zooms out, capturing the anticipation of arrival.",
    "negative_prompt": "Overexposure, static, blurred details, subtitles, still, overall gray, worst quality, low quality, JPEG compression residue, ugly, mutilated, redundant fingers, poorly painted hands, poorly painted faces, deformed, disfigured, deformed limbs, fused fingers, three legs",
    "aspect_ratio": "16:9",
    "lora_settings": [
      {
        "lora_type": "ghibli",
        "lora_strength": 1.0
      }
    ]
  }
}

Response Format

{
  "task_id": "task_12345",
  "status": "pending",
  "message": "Task submitted successfully",
  "estimated_time": "60-120 seconds"
}

Error Handling

Common Error Codes

400

Invalid LoRA settings or missing required fields

401

Invalid or missing API key

422

Unsupported LoRA type or invalid strength value

429

Rate limit exceeded

Rate Limits & Pricing

Rate Limits

• 10 requests per minute
• 100 requests per hour
• 1000 requests per day

Pricing

• txt2video-14b-lora: $0.28 per generation
• img2video-14b-lora: $0.28 per generation
• Processing time: 60-120 seconds