> For the complete documentation index, see [llms.txt](https://docs.pletor.ai/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://docs.pletor.ai/models/video-models/runway-aleph.md).

# Runway Aleph 2.0

Runway's flagship video editing model: edits real footage from a prompt, guided by keyframe images.

## TL;DR

Edit up to 30 seconds of footage without a VFX team. Swap backdrops, colorways, clothing, props or lighting and keep the rest of the scene. Show the exact change on a keyframe image and Aleph 2.0 carries it through the clip.

### Strengths for marketers

* **Edits that hold the scene**: change one thing while motion and framing stay as shot.
* **Keyframe-guided control**: up to 5 images on the clip's timeline show what a moment should look like.
* **Up to 30-second sources**, six times the old 5-second limit.
* **Keeps your resolution**: output matches the source, up to 1080p. Nothing is auto-cropped.
* **Expand to a new format**: pick a target aspect ratio and Aleph 2.0 generates the new edges.

### Ideal use cases

* **Product swaps**: a new colorway or SKU in existing lifestyle footage.
* **Localization and seasonality**: signage, weather, backdrop or time of day.
* **Campaign variations** from one shoot.
* **Format adaptation** by outpainting rather than cropping.

### Weaknesses

* Only edits existing video.
* Single continuous shots work best; a source with many cuts can be rejected.
* No audio generation or editing.

***

## How to use effectively

Short, direct prompts: an action word (replace, change, remove, add, re-light) and the change. When words aren't precise enough, edit a still of a frame and add it as a keyframe at that second. Keyframes you don't place are spread evenly across the clip.

### Example prompts

* "Replace the white sneakers with the red ones from the keyframe"
* "Change the backdrop to a snowy mountain street"
* "Re-light the scene for golden hour"

***

## Model parameters

* **Inputs**: video up to 30 seconds, 30 fps or lower, 16 MB or less, ratio between 1:2 and 2:1; prompt up to 1,000 characters; up to 5 keyframe images.
* **Output**: same resolution (up to 1080p) and duration as the source. **Auto** keeps the framing, or pick 16:9, 9:16, 1:1, 4:3, 3:4, 3:2, 2:3 or 21:9 to outpaint to that format.
