"Inference Engineering" is now available. Get your copy here
image generation

Qwen LogoWan 2.2 T2V

14B parameter text-to-video generation model from Wan AI.

Model details

View repository

Example usage

Generate a video from a text prompt. The response contains a base64-encoded video.

Input
1import requests
2import base64
3import os
4
5# Supported video sizes:
6# 720x1280, 1280x720, 480x832, 832x480,
7# 704x1280, 1280x704, 1024x704, 704x1024
8
9res = requests.post(
10    "https://model-{MODEL_ID}.api.baseten.co/production/predict",
11    headers={"Authorization": f"Api-Key {os.environ['BASETEN_API_KEY']}"},
12    json={
13        "prompt": "A cat playing piano in a jazz club, cinematic lighting",
14        "size": "1280x720",
15        "sampling_steps": 40,
16        "guide_scale": [3.0, 4.0],
17        "seed": 42,
18    },
19)
20
21# Save the base64-encoded video to an mp4 file
22video_data = base64.b64decode(res.json()["video"])
23with open("wan2_output.mp4", "wb") as f:
24    f.write(video_data)
25
JSON output
1{
2    "video": "AAAAIGZ0eXBpc29tAAACAGlzb21pc28yYXZj..."
3}

🔥 Trending models