AI Video Editor Auto Clip API Documentation

AI Video Editor Auto Clip API Documentation

Playground

AI Video Editor Auto Clip cuts long or raw footage down to a short highlight clip: it reviews one or several videos end to end, picks the moments that matter, follows your prompt for topic, style and length, and delivers a captioned clip of roughly 15 to 60 seconds. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

Auto Clip cuts long or raw footage down to one short clip, typically 15–60 seconds and never longer than 120. It is a clipping tool rather than a full-length editor: it decides what is worth keeping and returns a single short video. Upload one or several source videos, optionally describe the topic, style, or target length, and the editor reviews the footage, selects the most relevant moments, arranges them into a coherent sequence, and adds captions when verified speech is available.

Use Auto Clip when you want the AI to decide which parts of the footage should make the final cut. It is designed for turning long recordings, multiple clips, product footage, events, tutorials, interviews, or other raw material into a concise highlight or social-ready video without manually reviewing and assembling the timeline.

Why Choose This?

• Automatic video editing
  Turn raw footage into a finished short-form edit without manually selecting clips or arranging a timeline.

• Whole-source review
  Source footage is reviewed before moments are selected, helping the edit focus on relevant content rather than simply using the beginning of the video.

• Prompt-guided editing
  Describe the topic, editing direction, or desired length, such as a 30-second highlight of the finished result.

• Works with spoken and visual content
  Spoken footage is edited around complete sentences. When usable speech is unavailable, the editor can select content based on what is visible while retaining the original audio.

• Multi-video support
  Combine up to 8 source videos into a single finished edit.

• Flexible aspect ratios
  Create landscape, vertical, square, portrait, or ultrawide output for different platforms and publishing workflows.

• Captions under your control
  Word-by-word highlighted captions by default, plain static captions, or none at all with caption_style. Captions are generated only for verified speech and are repositioned when the source already contains burned-in subtitles.

Parameters

Output

outputs lists the edited MP4 first, followed by its SRT subtitle file. The subtitles are also embedded in the MP4 as a soft subtitle track (off by default), so players such as QuickTime, VLC and video editors can switch them on. The SRT file is returned with every caption_style, including none.

How to Use

1. Upload source videos — Provide between 1 and 8 videos to edit.
2. Describe the clip (optional) — Use prompt to specify the topic, style, focus, or desired output length.
3. Choose a caption style (optional) — Use caption_style to keep word-by-word highlighting, switch to plain captions, or turn captions off.
4. Choose aspect ratio (optional) — Select a landscape, vertical, square, or other supported format, or omit it to follow the source.
5. Choose language (optional) — Select the spoken language or keep auto for automatic detection.
6. Submit — Let the editor review the footage, select relevant moments, arrange the cut, and generate captions when applicable.
7. Retrieve the result — Download the completed edited video.

Pricing

Billing follows the combined length of every source video, per started hour.

Billing Rules

• $0.50 per started hour of combined source footage, with a 1-hour minimum.
• The length of the finished clip does not affect the price, and aspect_ratio, caption_style and language add no separate charge.

Best Use Cases

• Long recordings — Reduce lectures, streams, webinars, or walkthroughs to their key moments.
• Multi-clip footage — Combine several takes or angles of the same subject into one clip.
• Tutorials and projects — Show the key steps and the finished result rather than the whole process.
• Events and travel — Turn hours of footage into a short, shareable clip.
• Social formats — Produce vertical or square clips for short-form feeds.

Pro Tips

• Name the outcome you want in prompt, such as the key steps and the finished result, rather than only a topic.
• State a target length in the prompt when it matters; otherwise the clip length follows the content.
• With several videos, each one receives a share of the clip. When there are more videos than the length can carry, the best-matching ones are kept and played in source order.
• Footage without usable speech is selected by what is visible, in chronological order, so a process still ends on its result.
• Use 9:16 for vertical feeds and leave aspect_ratio unset to keep the original framing.

Notes

• videos is required and supports between 1 and 8 source videos.
• Maximum input size is 2 GB per video.
• Processing usually takes 4–10 minutes and depends only slightly on the source length.
• Output is one short clip: typically 15–35 seconds without a prompt, and at most 120 seconds.
• When a target length is specified in the prompt, the editor uses it to guide the output duration.
• Spoken content is edited around complete sentences when usable speech is available.
• Footage without usable speech is edited based on visible content in chronological order while retaining the original audio.
• Captions are generated only for speech that can be verified.
• When burned-in subtitles are detected in the source, generated captions are positioned to avoid covering them.
• When changing aspect ratio, the full source frame is retained and remaining canvas space is filled with a blurred copy of the footage.

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'
{
  "videos": [
    "https://interactive-examples.mdn.mozilla.net/media/cc0-videos/flower.mp4"
  ],
  "aspect_ratio": "16:9",
  "caption_style": "highlight",
  "language": "auto"
}
JSON
)

# 1. Submit the prediction.
SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  -X POST "https://api.wavespeed.ai/api/v3/wavespeed-ai/ai-video-editor/auto-clip" \
  -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
done

Parameters

Task Submission Parameters

Request Parameters

Response Parameters

Result Request Parameters

Result Response Parameters