Turn an Image into a Video

Use this guide to make a video whose first frame — or last frame — is a picture you give. By the end you will have submitted an image-to-video job with frame_images, and downloaded the clip.

Submitting POST /v1/videos starts a real video job and is billed in DA when the clip is delivered.

Before you start#

• An API key in OPENDUNES_API_KEY.
• A picture at a public HTTPS address that answers 200 with an image type — no cookies, no redirect through a web page. Or a picture you uploaded to Files.
• A video model that takes a frame. Video models are listed at GET /v1/videos/models; a model's supported_frame_images names the frames it takes — first_frame, last_frame, or both.

Step 1: Submit the job#

curl https://opendunes.com/v1/videos \
  -H "Authorization: Bearer $OPENDUNES_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/veo-3.1-lite",
    "prompt": "Sand slowly rippling in a gentle wind, golden hour light",
    "duration": 4,
    "frame_images": [
      {
        "type": "image_url",
        "image_url": { "url": "https://your-cdn.example/dunes-still.jpg" },
        "frame_type": "first_frame"
      }
    ]
  }'

The answer is 202, with the job's id, its polling_url and its status. prompt may be left out when the picture is the whole input — for a model that needs one, the model's provider refuses and its words come back.

At most two items. Every other key — resolution, aspect_ratio, seed — travels to the model as you send it; a value the model does not take is refused with the values it does.

Step 2: Add a last frame for a transition#

With both frames, the model moves from one picture to the other:

"frame_images": [
  { "type": "image_url", "image_url": { "url": "https://your-cdn.example/first.jpg" }, "frame_type": "first_frame" },
  { "type": "image_url", "image_url": { "url": "https://your-cdn.example/last.jpg" }, "frame_type": "last_frame" }
]

Step 3: Poll and download#

Poll GET /v1/videos/{id} until status is completed, then fetch the clip from GET /v1/videos/{id}/content. Polling is free. Or pass a callback_url and be told when it ends — see Get Video Results with Webhooks. The whole loop is in Generate and Download a Video from Text.

Frames or references#

frame_images pins an exact frame. To guide a video's subject or style without pinning a frame, use input_references — see Guide a Video with Reference Images. Sent together, the request is treated as image-to-video.