ltx2

Generates short video clips from text prompts or images using the LTX-2.3 22B model on Modal.

Updated Sep 1, 2026
One-click install
npx skills add https://github.com/zamansepeti43/c-rak-agent --skill ltx2-zamansepeti43
Or copy as Structured Prompt for Agent
Please help me install this Agent Skill.
Skill: ltx2
Source: https://github.com/zamansepeti43/c-rak-agent/tree/main/video-engine/.agents/skills/ltx2
Command: npx skills add https://github.com/zamansepeti43/c-rak-agent --skill ltx2-zamansepeti43

SYSTEM DOCUMENTATION & REQUIREMENTS

What problem does it solve? Creating video content like b-roll, animated backgrounds, or motion clips traditionally requires filming, stock footage licensing, or complex animation tools. This Skill generates ~5 second video clips directly from text prompts or still images using the LTX-2.3 22B DiT model, removing the need for manual video production. ## Core Features & Use Cases - Text-to-Video Generation: Create cinematic clips from descriptive prompts with control over resolution, frame count, quality, and seed. - Image-to-Video Animation: Animate still images such as headshots, slides, or product photos with subtle motion. - Production Workflows: Generate b-roll, animated slide backgrounds, branded intro/outro backgrounds, and combine outputs with upscaling, Remotion compositions, or voiceover tools. - Use Case: A video producer needs atmospheric b-roll of a European city at golden hour for a narrated segment. They run a single command with a cinematographic prompt and receive a 5-second H.264 MP4 clip ready to drop into the edit. ## Quick Start Ask the agent to generate a 5-second video clip of a sunset over the ocean with cinematic golden light and save it as sunset.mp4.

Frequently Asked Questions about ltx2

High-intent search queries and answers about installing and using this skill.

FAQPage Schema
How do I generate a video from a text prompt with LTX-2?

Run the ltx2.py tool with a --prompt describing the scene and an --output path, for example a cinematic sunset description saved as an MP4. You can adjust width, height, frame count, quality, and seed for control over the result.

How to animate a still image into a video clip?

Use image-to-video mode by passing --input with your image path alongside a motion-focused prompt like gentle camera drift or subtle head movement. The model animates the still image into a short MP4 clip at 24fps.

What is the maximum video length LTX-2 can generate?

The maximum practical duration is about 8 seconds at 193 frames, with 121 frames (~5 seconds) as the default. Frame counts must satisfy (n-1) % 8 == 0, and longer content requires stitching multiple clips together.

Can LTX-2 generate readable text inside videos?

No, the model cannot reliably render readable text in video output. For titles, captions, or UI text, generate the background clip with LTX-2 and overlay text using Remotion compositions instead.

Why does my generated video contain unwanted logos or watermarks?

Roughly 30% of generations may include logos or text artifacts inherited from training data. Re-run the generation with a different --seed value to get a clean result.

What GPU and cost does LTX-2 video generation require?

Generation runs on an A100-80GB GPU via Modal at roughly $4.68 per hour, costing about $0.20-0.25 per 5-second clip. Expect around 2.5 minutes of inference per clip plus a 60-90 second cold start for loading weights.