AI Video Editing API — Restyle, Extend, Reframe, and Clip

Six Ways to Edit an Existing Video

Instruction edits (Runway Aleph, Seedance 2.5 Video Edit, Wan 2.2 Edit Video) restyle or transform a video from a text prompt while preserving its motion and structure. Extension (Seedance 2.5 Video Extend, PixVerse V6 Extend) continues a clip past its original end point, driven by its last frame plus a prompt. Reframing (Luma Flash Reframe, Autocrop) changes a video's aspect ratio intelligently, rather than cropping content out. Motion control (Kling v3.0 Pro Motion Control) applies precise camera-path direction to existing footage. Captions and clipping (AI Captions, AI Clipping) turn long-form footage into captioned, highlight-driven short clips. Utilities (Video Combiner, HeyGen Video Translate, AI Video Watermark Remover) stitch clips together, dub a video into another language, or strip a Sora 2 watermark, logo, or unwanted text from an existing clip.

Muapi exposes 70+ video-to-video models — Runway, Kling, Seedance, Wan, PixVerse, Luma, HeyGen, and more — through the same unified REST pattern: one API key, one request/poll flow, pay-per-generation pricing with no subscription.

Related Video-to-Video Tasks

A few adjacent video-to-video capabilities have their own dedicated pages rather than being duplicated here: face swap (replacing a face throughout a clip), background removal (frame-accurate matting), lipsync (re-syncing mouth movements to new audio), and video upscaling (resolution enhancement).

Prompt Structure and Input Handling

• Instruction-edit prompts should describe the change, not the whole scene — the source video already fixes subject identity, composition, and motion; the prompt drives what changes on top of that.
• Extension models use the source video's last frame as the seed for the new continuation — a stable, clearly composed final frame in the source clip produces a smoother extension than one mid-motion.
• Reframe and Autocrop work on a specific segment, not necessarily the whole clip — Autocrop explicitly takes a start_time/end_time range plus a target aspect_ratio, using AI subject tracking to keep the subject in frame.
• AI Clipping returns highlight segments, not a single output — request a specific num_highlights, or set return_coordinates_only to get timestamps/crop coordinates back instead of rendered clips, useful for a custom editing pipeline.

How to Call a Video Editing Model

1. Host the source video. Upload via POST /api/v1/upload_file, or use any publicly reachable URL.
2. Submit the request. POST /api/v1/{model-slug} with video_url and the fields that model supports (prompt, aspect_ratio, language, etc.).
3. Poll for completion. Check GET /api/v1/predictions/{request_id}/result until status is completed, then download the result.

curl -X POST https://api.muapi.ai/api/v1/runway-aleph-v2v \
  -H "Content-Type: application/json" \
  -H "x-api-key: YOUR_API_KEY" \
  -d '{
    "video_url": "https://example.com/source-clip.mp4",
    "prompt": "restyle as a golden-hour cinematic look",
    "aspect_ratio": "16:9"
  }'

import requests

response = requests.post(
    "https://api.muapi.ai/api/v1/runway-aleph-v2v",
    headers={"x-api-key": "YOUR_API_KEY"},
    json={
        "video_url": "https://example.com/source-clip.mp4",
        "prompt": "restyle as a golden-hour cinematic look",
        "aspect_ratio": "16:9",
    },
)
request_id = response.json()["request_id"]

result = requests.get(
    f"https://api.muapi.ai/api/v1/predictions/{request_id}/result",
    headers={"x-api-key": "YOUR_API_KEY"},
)
print(result.json())

Frequently Asked Questions