Back to Nodes
Lightricks LTX 2 Retake

Lightricks LTX 2 Retake

Official

AI-powered video retake tool for editing specific shot sections, reshaping dialogue, action, and camera angles within existing footage.

Nodespell AI
AI / Video / Lightricks

AI-powered video retake tool for editing specific shot sections, reshaping dialogue, action, and camera angles within existing footage.

Overview

LTX Retake is a creative AI feature designed to give directors and creators precise control over individual shots in a video. It enables editing specific segments by rephrasing dialogue while maintaining voice consistency, altering actions, changing camera motion, or redirecting story elements without losing the rest of the scene’s continuity. The model regenerates selected time ranges within a video using input prompts that describe the desired edits, producing output that blends naturally with unedited footage.

Strengths / What it does well

  • Enables granular, directable edits within shots, such as dialogue rephrasing and reframing.
  • Maintains motion, sound, and visual continuity outside the edited segment.
  • Supports iterative creative adjustments to pacing, tone, and story direction.
  • Facilitates immediate feedback implementation for filmmakers and content creators.

Limitations

  • Maximum input video length supported is about 20 seconds.
  • Output resolution is currently limited to 1080p.
  • Minimum editable segment is around 2–2.5 seconds, with a maximum segment length of 16 seconds.

Best use cases

  • Creative professionals seeking to fine-tune specific moments in a shot without reshooting.
  • Brands and directors needing fast revisions in video content.
  • Any user looking to reshape intent or emotion in pre-existing video while preserving overall scene integrity.
Inputs (2)

Video

String

Input video to edit

Min: 0Max: 100

Prompt

String

Text prompt describing the edit

Multi InputMin: 0Max: 100
Parameters (4)

Mode

String

Mode of operation

Default: replace_video

Prompt

String

Text prompt describing the edit

Default:

Duration

Number

Duration of the edited section in seconds

Default: 6

Start Time

Number

Start time of the edit in seconds

Default: 0
Outputs (1)

Output

Inferred

Output

Nodespell

Nodespell

📍 London

Building the future. Join us!

Type

Node

Status

Official

Package

Nodespell AI

Category

AI / Video / Lightricks

Input

VideoText

Output

Video

Keywords

Video EditLength Control
Use in Workflow