Products

Text to Clip

Describe a moment and get a still image, a 5 second video clip, and a cover frame from one call.

Text to Clip (text-to-clip) turns a one-line idea into three files: a still image of the scene, a 5 second clip that animates it, and a cover frame taken from the end of the clip. It’s the quickstart’s image-then-video flow in a single request.

curl

curl https://api.windpaint.ai/v1/workflows/products/text-to-clip/runs \
  -H "Authorization: Bearer $WINDPAINT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"inputs": {"idea": "a paper boat drifting down a rainy street", "aspect": "9:16"}}'

Inputs

ideatextrequired

What should happen in the clip, in plain words. It’s used as the prompt for both the still and the motion, so describe the scene and what moves: “a paper boat drifting down a rainy street”, “steam rising from a coffee cup on a windowsill at sunrise”.

aspectchoicedefault:16:9

Aspect ratio of every output. One of 16:9 (landscape) or 9:16 (vertical). The API lists this input as required, but it has a default, so you can leave it out.

Steps

Step	Kind	What it does	Credits
`still`	`image.generate` with `z-image-turbo`	Generates the scene from your idea, styled as a cinematic still with natural light, at the `1k` tier in your aspect ratio.	0.08
`frame`	`image.resize` op	Crops and resizes the still to the size the video model expects.	free
`clip`	`video.generate` with `wan-2.2-i2v`	Animates the resized still into a 5 second, 480p clip, using your idea as the motion prompt.	4
`cover`	`video.extract_frame` op	Takes the last frame of the clip as a PNG.	free

The steps run in that order, since each one uses the previous step’s output. Most of the time is the clip step, which takes a few minutes.

Outputs

Output	Kind	Description
`video`	video (MP4)	The 5 second, 480p clip.
`still`	image	The generated still, before resizing.
`cover`	image (PNG)	The clip’s last frame, for use as a thumbnail or the start of a follow-up clip.

All three are saved as assets in the run’s project. The resized frame is saved too and appears in that step’s outputs, but it isn’t one of the run’s named outputs.

Price

A run costs 4.08 credits at current prices: 0.08 for the still and 4 for the clip. Both aspect ratios cost the same. Check before you run:

curl

curl https://api.windpaint.ai/v1/workflows/products/text-to-clip/estimate \
  -H "Authorization: Bearer $WINDPAINT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"inputs": {"idea": "a paper boat drifting down a rainy street"}}'

If the clip step fails, you still pay for the still (0.08 credits) and nothing for the clip. See Billing for how runs are charged.

Full example

Start the run, poll until it finishes, then download the three outputs.

curl

# Start the run and keep its id
run_id=$(curl -s https://api.windpaint.ai/v1/workflows/products/text-to-clip/runs \
  -H "Authorization: Bearer $WINDPAINT_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"inputs": {"idea": "steam rising from a coffee cup on a windowsill at sunrise", "aspect": "16:9"}}' \
  | jq -r '.data.id')

# Poll every 10 seconds until the run is terminal
while :; do
  run=$(curl -s "https://api.windpaint.ai/v1/workflows/runs/$run_id" \
    -H "Authorization: Bearer $WINDPAINT_API_KEY")
  status=$(echo "$run" | jq -r '.data.status')
  echo "$status"
  case "$status" in completed|failed|canceled) break ;; esac
  sleep 10
done

# Download each output (follow the redirect to the signed URL)
for name in video still cover; do
  url=$(echo "$run" | jq -r ".data.outputs.$name[0].url")
  ext=$(echo "$run" | jq -r ".data.outputs.$name[0].content_type" | cut -d/ -f2)
  curl -sL "$url" -H "Authorization: Bearer $WINDPAINT_API_KEY" -o "$name.$ext"
done

A completed run looks like this (steps trimmed):

{
  "data": {
    "id": "7c1e9a40-3b2d-4f6a-8e15-0d9c4b7a2f63",
    "status": "completed",
    "product": "text-to-clip",
    "product_version": 1,
    "inputs": { "idea": "steam rising from a coffee cup on a windowsill at sunrise", "aspect": "16:9" },
    "steps": [ ... ],
    "outputs": {
      "video": [{ "id": "e2a7...", "url": "https://api.windpaint.ai/v1/generation/assets/e2a7.../content", "content_type": "video/mp4", "duration_s": 5.0 }],
      "still": [{ "id": "d77a...", "url": "https://api.windpaint.ai/v1/generation/assets/d77a.../content", "content_type": "image/png", "width": 1024, "height": 576 }],
      "cover": [{ "id": "91fc...", "url": "https://api.windpaint.ai/v1/generation/assets/91fc.../content", "content_type": "image/png" }]
    },
    "credits": { "estimate": "4.08", "actual": "4.08" },
    "error": null,
    "finished_at": "2026-10-04T14:13:41.902Z"
  }
}

If status is failed, error names the step that failed and why, and that step’s error has the detail. The run’s outputs stay empty, but the outputs of steps that completed (the still, for example) are in steps[].outputs and in your project’s assets.

Tips

Write the idea as a scene with motion. The same text drives the still and the animation, so “a paper boat drifting down a rainy street” works better than “paper boat”.
Chain clips with the cover. The cover is the clip’s last frame. Pass its asset id as start_frame to video.generate to continue the shot. See Video.
Use 9:16 for vertical feeds and 16:9 for everything else. Both cost the same.

Other products

The catalog lists other products too. They depend on capabilities that don’t have a model yet, so they report available: false and can’t be run. They’ll start working when those models ship, without any change on your side. See Availability.