Kwaivgi Kling V4.0 Flash Reference To Video API Documentation
Playground
Try it on WaveSpeedAI!Kling V4.0 Flash Reference-to-Video generates a 3 to 20 second 720p clip guided by up to 10 reference images and 5 reference videos, with synchronized audio; Flash is the fast, lower-cost tier of Kling 4.0. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.
Features
Kling V4.0 Flash Reference-to-Video generates a 720p video guided by reference images and videos, with synchronized audio. Flash is the fast, lower-cost tier of Kling 4.0: 720p clips of 3 to 20 seconds with synchronized audio, for quick iteration and volume work.
Why Choose This?
- Consistent subjects — Up to 10 reference images for characters, products or styles.
- Motion references — Up to 5 reference videos (together up to 20 seconds) for motion, camera and pacing.
- Keep the original sound — Optionally keep a reference video’s own audio.
- Fast iteration — Flash returns clips quickly.
Parameters
| Parameter | Required | Description |
|---|---|---|
| prompt | Yes | Describe the video. Refer to reference images as @Image1, @Image2, … and to reference videos as @Video1, @Video2, … in array order. |
| images | No | Up to 10 reference images. |
| videos | No | Up to 5 reference videos, together used for up to 20 seconds. |
| keep_original_sound | No | Keep the reference video’s own sound instead of generating new audio. Requires videos. Default false. |
| duration | No | 3 to 20 seconds, default 5. |
| aspect_ratio | No | auto (default, follows the references), 16:9, 9:16, 1:1 or 21:9. |
Pricing
| Inputs | Price per second of output |
|---|---|
| Reference images only | $0.094 |
| With reference videos | $0.126 |
A 5 second clip costs $0.47 with images only, $0.63 with reference videos.
Related Models
- Kling V4.0 Flash Text-to-Video — Generate a clip from a text prompt.
- Kling V4.0 Flash Image-to-Video — Animate a first-frame image.
- Kling V4.0 Flash Video Edit — Edit an existing video with an instruction.
- Kling V4.0 — The full Kling 4.0: 1080p, clips up to 30 seconds, keyframes and reference audio.
Authentication
For authentication details, please refer to the Authentication Guide.
API Endpoints
Submit Task & Query Result
set -euo pipefail
export WAVESPEED_API_KEY="your-api-key"
REQUEST_BODY=$(cat <<'JSON'
{
"prompt": "A cinematic ocean wave at sunrise, highly detailed",
"keep_original_sound": false,
"duration": 5,
"aspect_ratio": "auto"
}
JSON
)
# 1. Submit the prediction once. Never retry this request automatically:
# a retry after an unclear failure can create a second, billed task.
CHECK='The task may already exist. Check History or the predictions list before submitting again.'
if ! SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
--connect-timeout 10 --max-time 60 \
-X POST "https://api.wavespeed.ai/api/v3/kwaivgi/kling-v4.0-flash/reference-to-video" \
-H "Authorization: Bearer ${WAVESPEED_API_KEY}" \
-H "Content-Type: application/json" \
-d "${REQUEST_BODY}"); then
printf 'Submit failed. %s\n' "${CHECK}" >&2
exit 1
fi
TASK=$(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. %s\n' "${CHECK}" >&2
exit 1
fi
RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"
# 2. Poll until the prediction reaches a final status.
BODY_FILE=$(mktemp)
trap 'rm -f "${BODY_FILE}"' EXIT
while true; do
# Transient errors (429, 5xx) are retried, honoring Retry-After.
HTTP_CODE=$(curl --silent --show-error --fail-with-body \
--connect-timeout 10 --max-time 60 --retry 5 --retry-delay 2 \
--output "${BODY_FILE}" --write-out '%{http_code}' \
"${RESULT_URL}" \
-H "Authorization: Bearer ${WAVESPEED_API_KEY}" || true)
case "${HTTP_CODE}" in
2??) ;;
000|429|5??) # dropped connection or temporary error: keep polling
printf 'Task %s: HTTP %s, polling again in 10s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
sleep 10; continue ;;
*) printf 'Task %s: HTTP %s\n' "${PREDICTION_ID}" "${HTTP_CODE}" >&2
cat "${BODY_FILE}" >&2; exit 1 ;;
esac
RESULT=$(jq 'if type == "object" and has("data") then .data else . end' "${BODY_FILE}")
STATUS=$(printf '%s' "${RESULT}" | jq -r '.status // empty')
case "${STATUS}" in
completed) printf '%s\n' "${RESULT}" | jq '.outputs'; break ;;
failed|cancelled|timeout|deleted)
printf 'Task %s ended with status %s\n' "${PREDICTION_ID}" "${STATUS}" >&2
printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
*) sleep 2 ;;
esac
doneRetry temporary errors only while polling, never the submit request. Errors, retries and timeouts
Parameters
Task Submission Parameters
Request Parameters
| Parameter | Type | Required | Default | Range | Description |
|---|---|---|---|---|---|
| prompt | string | Yes | - | The positive prompt for the generation. Refer to reference images as @Image1, @Image2, ... (in array order) and to reference videos as @Video1, @Video2, ... | |
| images | array<string> | No | - | 0 ~ 10 items | Reference images (up to 10). |
| videos | array<string> | No | - | 0 ~ 5 items | Reference videos (up to 5). They guide motion, camera and pacing; together they are used for up to 20 seconds. |
| keep_original_sound | boolean | No | false | - | Keep the reference video's own sound instead of generating new audio. Requires videos. |
| duration | integer | No | 5 | 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19, 20 | The duration of the generated video in seconds (3-20). |
| aspect_ratio | string | No | auto | auto, 16:9, 9:16, 1:1, 21:9 | The aspect ratio of the generated video; auto follows the reference images or video. |
Every model also accepts the global request options enable_sync_mode and enable_base64_output; see their pages for details.
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 |