Text to Video

Example

import requests

DEVUP_API_KEY = "$DEVUP_API_KEY"
MODEL = "Wan-AI/Wan2.2-T2V-A14B"

response = requests.post(
    "https://api.devupai.com/v1/video/generations",
    headers={
        "Authorization": f"Bearer {DEVUP_API_KEY}",
        "Content-Type": "application/json",
    },
    json={
        "model": MODEL,
        "prompt": "A serene mountain lake at sunrise, with mist rising from the water and pine trees reflected on the surface.",
    },
)

result = response.json()
# result["data"][0]["url"] contains the signed URL to the generated video
video_url = result["data"][0]["url"]
video_response = requests.get(video_url)

with open("output.mp4", "wb") as f:
    f.write(video_response.content)

Response format options

The response_format parameter controls how the generated video is returned. It accepts "url" (default) or "b64_json". Requesting "url" returns a signed proxy URL in data[].url. Requesting "b64_json" returns the video bytes Base64-encoded inline in data[].b64_json and writes nothing to storage. Any other value is rejected with HTTP 400.

Response format

{
  "data": [
    {
      "url": "https://api.devupai.com/v1/images/proxy?d=<b64url>&s=<hmac>"
    }
  ],
  "_devup": {
    "cost_dzd": 0.0,
    "balance_dzd": 0.0
  }
}

{
  "data": [
    {
      "b64_json": "<base64>"
    }
  ],
  "_devup": {
    "cost_dzd": 0.0,
    "balance_dzd": 0.0
  }
}

When "b64_json" is requested and the generated video exceeds the 20 MB inline size limit, the response automatically falls back to a signed URL in data[].url and includes a note field inside _devup explaining why. This is an automated delivery fallback, not an error, and the generation request is billed normally.

URL expiry

The generated video URL in data[].url is signed and expires 300 seconds after the response is generated. After expiry, the proxy returns HTTP 403 Forbidden. Download the file rather than storing the URL. This expiration applies only to the "url" response; callers who do not want an expiring URL can request "b64_json" instead.

Generation time

Video generation is synchronous and a single request may hold the connection open for several minutes. Clients should set a generous timeout when making requests.

Tips for good prompts

• Be descriptive about the scene, lighting, and motion
• Specify the camera movement if relevant (e.g. "slow pan", "aerial shot", "close-up")
• Keep prompts focused — overly complex prompts can produce inconsistent results
• Use the negative prompt parameter (if supported) to exclude unwanted elements

Available models

Browse all text-to-video models.