Video Cleaner API Documentation
Playground
Try it on WaveSpeedAI!Video Cleaner removes people, vehicles and other moving subjects from a video and rebuilds the background behind them, producing a clean plate of the same shot with the camera motion intact. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.
Features
Get a clean plate from any shot. Video Cleaner removes people, pedestrians, vehicles and other moving subjects from a video and rebuilds the background behind them, keeping the camera movement, the static scene and the lighting exactly as they were. No masks to draw: upload the clip and the moving subjects are gone.
What It Does
-
Removes moving subjects People, crowds, cars, bikes and other dynamic elements are taken out across the whole frame.
-
Rebuilds the background Architecture, ground markings, foliage and street furniture are reconstructed where the subjects used to be, frame-consistent across the clip.
-
Keeps the shot Same framing, same camera motion, same length. Works with static and moving cameras.
Parameters
| Parameter | Required | Description |
|---|---|---|
| video | Yes | Input video (URL or upload). Up to 2 minutes is processed. |
| prompt | No | Extra objects that must also disappear, comma-separated (e.g. “bicycle, dog”) |
| target_resolution | No | Output resolution: 480p, 720p, 1080p (default: 720p). Aspect ratio is preserved. |
| seed | No | Seed for reproducible results |
Pro Tips
- Removal works best when the subject leaves enough background visible around it. A subject that fills the centre of the frame (a first-person torso, a car filling the shot) leaves too little to rebuild.
- If a specific object survives, name it in
prompt(e.g. “bicycle, handlebars”). - For advertising plates, run the clip once, then composite your product onto the clean result.
Limits and Performance
- Max clip length per job: 2 minutes (longer input is truncated)
- Processing speed: roughly 8 seconds of wall time per second of video at 720p, about 20 seconds per second at 1080p
Pricing
> Billing duration: Input media duration is rounded up to whole seconds. The minimum billed duration is 3 seconds; shorter inputs are billed as 3 seconds. Input beyond 2 minutes is not processed and not billed.
Per-second billing with a 3-second minimum. The table below lists prices per 5 seconds for easy comparison.
| Output Resolution | Cost per 5 seconds |
|---|---|
| 480p | $0.10 |
| 720p | $0.20 |
| 1080p | $0.50 |
Billing Rules
- Minimum charge: 3 seconds
- Per-second rate = (price per 5 seconds) ÷ 5
- Billed duration = video length in seconds (rounded up), with a 3-second minimum, capped at 120 seconds
- Total cost = billed duration × per-second rate (by output resolution)
Notes
- Video is the only required field.
- Maximum processed video length per job: 2 minutes.
- Ensure video URLs are publicly accessible.
Authentication
For authentication details, please refer to the Authentication Guide.
API Endpoints
Submit Task & Query Result
set -euo pipefail
export WAVESPEED_API_KEY="your-api-key"
REQUEST_BODY=$(cat <<'JSON'
{
"video": "https://interactive-examples.mdn.mozilla.net/media/cc0-videos/flower.mp4",
"target_resolution": "720p"
}
JSON
)
# 1. Submit the prediction.
SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
-X POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/video-cleaner" \
-H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
-H "Content-Type: application/json" \
-d "${REQUEST_BODY}")
TASK=$(printf '%s' "${SUBMIT_RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
PREDICTION_ID=$(printf '%s' "${TASK}" | jq -r '.id // empty')
if [ -z "${PREDICTION_ID}" ]; then
printf 'Submission response did not contain a prediction id
' >&2
exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"
# 2. Poll until the prediction finishes.
while true; do
RESPONSE=$(curl --silent --show-error --fail-with-body \
"${RESULT_URL}" \
-H "Authorization: Bearer ${WAVESPEED_API_KEY}")
RESULT=$(printf '%s' "${RESPONSE}" | jq 'if type == "object" and has("data") then .data else . end')
STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')
case "${STATUS}" in
completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
failed|cancelled|timeout|deleted) printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
*) sleep 2 ;;
esac
doneParameters
Task Submission Parameters
Request Parameters
| Parameter | Type | Required | Default | Range | Description |
|---|---|---|---|---|---|
| video | string | Yes | - | The video to clean. Works best when the subjects to remove leave enough background visible around them; up to 2 minutes is processed (longer input is truncated). | |
| prompt | string | No | - | Optional. Extra objects that must also disappear, comma-separated (for example "bicycle, dog"). People and vehicles are removed by default. | |
| target_resolution | string | No | 720p | 480p, 720p, 1080p | Output resolution. The input aspect ratio is preserved; output is never larger than the input. |
| seed | integer | No | - | - | The seed for random number generation. Omit for a random seed. |
Response Parameters
| Parameter | Type | Description |
|---|---|---|
| code | integer | HTTP status code (e.g., 200 for success) |
| message | string | Status message (e.g., “success”) |
| data.id | string | Unique identifier for the prediction, Task Id |
| data.model | string | Model ID used for the prediction |
| data.outputs | array | Output values, usually URL strings; some models return text strings or structured result objects (empty when status is not completed) |
| data.urls | object | Object containing related API endpoints |
| data.status | string | Task status. completed is successful; failed, cancelled, timeout, and deleted are failure terminal statuses. |
| data.created_at | string | ISO timestamp of when the request was created (e.g., “2023-04-01T12:34:56.789Z”) |
| data.error | string | Error message (empty if no error occurred) |
| data.timings | object | Object containing timing details |
| data.timings.inference | integer | Inference time in milliseconds |
Result Request Parameters
| Parameter | Type | Required | Default | Description |
|---|---|---|---|---|
| id | string | Yes | - | Task ID |
Result Response Parameters
| Parameter | Type | Description |
|---|---|---|
| code | integer | HTTP status code (e.g., 200 for success) |
| message | string | Status message (e.g., “success”) |
| data | object | The prediction data object containing all details |
| data.id | string | Unique identifier for the prediction |
| data.model | string | Model ID used for the prediction |
| data.outputs | array<string | object> | Array of generated outputs (empty when status is not completed). Items are usually URL strings, but may be text strings or structured result objects, depending on the model. |
| data.urls | object | Object containing related API endpoints |
| data.status | string | Status: completed is successful; failed, cancelled, timeout, and deleted are failure terminal statuses |
| data.created_at | string | ISO timestamp of when the request was created |
| data.error | string | Error message (empty if no error occurred) |
| data.timings | object | Object containing timing details |
| data.timings.inference | integer | Inference time in milliseconds |