Runwayml Aleph 2

Runwayml Aleph 2

Playground

Runway Aleph 2 Video Edit is an in-context video editing model for precise prompt-based edits, multi-shot consistency, and optional keyframe guidance, supporting 2-30 second input videos and up to 5 keyframes for controlled video modification workflows. Ready-to-use REST inference API, best performance, no coldstarts, affordable pricing.

Features

Runway Aleph 2 edits videos with natural-language instructions while preserving scene consistency. Upload a source video, describe the edit you want, and optionally provide keyframe images to guide specific moments in the edited result.

Why Choose This?

• Natural-language video editing
  Edit an existing video by describing the desired change in plain text.

• Scene-consistent edits
  Apply visual changes while preserving the overall structure and continuity of the source video.

• Multi-shot editing support
  Work with videos that contain multiple shots while maintaining coherent edit behavior.

• Optional keyframe guidance
  Provide up to 5 reference keyframes to guide the beginning, ending, or specific timestamps of the edited video.

• Seed control
  Use a seed when you need more reproducible edit results.

Parameters

How to Use

1. Upload your video — Provide a 2–30 second source video under 16 MB.
2. Write your edit prompt — Describe the visual change you want to apply.
3. Add keyframes optional — Upload up to 5 keyframe images when you need stronger guidance for specific moments.
4. Set keyframe positions optional — Use first, last, or timestamps in seconds to place each keyframe.
5. Set seed optional — Use a fixed seed when you want more reproducible results.
6. Submit — Generate the edited video.

Pricing

Pricing is $1.85 per 5 seconds of source video duration, prorated from 2 to 30 seconds. This is equivalent to $0.37 per second.

Best Use Cases

• Prompt-based video editing — Apply natural-language edits to existing clips.
• Scene restyling — Change mood, lighting, atmosphere, or visual direction while preserving scene continuity.
• Keyframe-guided edits — Guide the first frame, last frame, or specific timestamps with reference images.
• Multi-shot video edits — Edit clips that contain multiple shots while keeping the result coherent.
• Creative iteration — Test different edit directions from the same source video.

Pro Tips

• Use a clear source video within the supported 2–30 second range.
• Keep the edit prompt focused on the specific change you want.
• Use keyframe images when the edit needs stronger visual guidance at specific moments.
• Make sure each keyframe image has a matching keyframe position.
• Use first or last for simple start/end guidance.
• Use timestamp positions when a keyframe should guide a specific moment in the video.
• Use a fixed seed when comparing prompt or keyframe variations.

Note

Keyframe image and position arrays must contain the same number of entries.

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",
  "video": "https://interactive-examples.mdn.mozilla.net/media/cc0-videos/flower.mp4",
  "keyframe_images": [],
  "keyframe_positions": []
}
JSON
)

# 1. Submit the prediction.
SUBMIT_RESPONSE=$(curl --silent --show-error --fail-with-body \
  -X POST "https://api.wavespeed.ai/api/v3/runwayml/aleph-2" \
  -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 "${PREDICTION_ID}" ]; then
  printf 'Submission response did not contain a prediction id
' >&2
  exit 1
fi
RESULT_URL=$(printf '%s' "${TASK}" | jq -r '.urls.get // empty')
if [ -z "${RESULT_URL}" ]; then RESULT_URL="https://api.wavespeed.ai/api/v3/predictions/${PREDICTION_ID}/result"; fi

# 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) printf '%s\n' "${RESULT}" | jq . >&2; exit 1 ;;
    created|processing) sleep 2 ;;
    *) printf 'Unexpected status: %s
' "${STATUS}" >&2; exit 1 ;;
  esac
done

Parameters

Task Submission Parameters

Request Parameters

Response Parameters

Result Request Parameters

Result Response Parameters