How long does LTX-2 video generation take?

Most LTX-2 generations finish in about 30 seconds to a few minutes, depending on resolution, duration and quality mode. Fast mode at 1080p is quickest; Ultra mode at 4K with audio takes the longest. Your task runs in the background — you can watch its status on the My Tasks page and download the video the moment it completes.

How do I do image-to-video with LTX-2?

Switch the generator to Image-to-Video mode, upload a JPG or PNG, and optionally add a short prompt describing the motion you want (for example 'gentle wind, camera slowly zooms in'). LTX-2 animates the image while keeping its composition and style intact. High-resolution source images with a clear subject produce the best results.

Why does my video look different from my prompt?

Usually the prompt is too vague — 'a dog running' leaves the subject, setting, lighting and camera entirely up to the model. Rewrite it with specifics: subject details, one clear action, a camera move and lighting. If the motion or detail still falls short, re-run the same prompt in Pro mode, which renders with noticeably higher fidelity than Fast.

Can I use LTX-2 for free?

Yes — new accounts get free starter credits, enough to generate your first videos at no cost. After that you can buy credit packs or subscribe for monthly credits. Fast mode at 1.0x credit cost keeps experimentation cheap.

What's the best mode for beginners?

Start with Fast mode at 1080p and 6 seconds. It costs the fewest credits per generation, so you can iterate on prompts freely. Once a prompt produces the motion you want, re-render it in Pro (1.5x) for final quality, or Ultra (2.0x) if you need native 4K with synchronized audio.

Beginner-friendly LTX-2 tutorial — updated for LTX-2.3

How to Use LTX-2

LTX-2 is Lightricks' open-source AI video model — the latest release, LTX-2.3, generates native 4K clips with synchronized audio from a text prompt or a single image. This tutorial shows you exactly how to generate your first video online at LTX2.info in under a minute — no installation, no GPU, no code. New to the model itself? Read the LTX-2 model overview first, then come back and follow the five steps below.

Open the generator See LTX-2 Pro mode

Generate an LTX-2 video in 5 steps

From blank page to downloadable 4K clip — the whole flow takes less than a minute of your time.

Sign up and claim your free credits

Create an account with Google or GitHub — it takes about ten seconds. Every new account receives free starter credits, which is enough to generate your first LTX-2 videos without paying anything. Your credits appear in the top bar as soon as you log in.

Choose Text-to-Video or Image-to-Video

Pick Text-to-Video to generate a clip from a written description alone. Pick Image-to-Video to animate a still picture: upload a JPG or PNG and LTX-2 brings it to life while preserving its composition and style. Image mode is ideal for product shots, artwork and portraits.

Write your prompt

Describe the subject, its motion, the camera movement, the lighting and the visual style — in that order works well. For example: 'A red vintage car driving along a coastal road at golden hour, drone shot following from behind, cinematic'. Keep one scene per clip: LTX-2 clips are 6–10 seconds, so a single continuous action reads far better than a multi-shot storyline.

Pick your settings

Choose resolution (1080p, 2K or 4K), duration (6, 8 or 10 seconds) and quality mode. Fast is best for drafts at 1.0x credit cost; Pro delivers final-quality output at 1.5x; Ultra forces native 4K with synchronized audio at 2.0x. Toggle audio on if you want LTX-2 to generate matching sound in the same pass.

Generate, preview and download

Hit Generate and watch the status in My Tasks — most clips finish within a couple of minutes. Preview the result in the browser, then download the MP4 with one click. Everything you generate is commercially usable, so you can publish it in ads, social posts and client work.

LTX-2 prompt tips

Small changes in wording make a big difference in output quality. These habits get consistently better results.

Be specific, not vague

'A golden retriever puppy running through tall grass at sunset, slow motion, shallow depth of field' beats 'a dog running' every time. Detail gives the model something concrete to render.

Direct the camera

Name a camera move: 'slow dolly in', 'drone shot pulling back', 'handheld tracking shot', 'static wide shot'. LTX-2 follows camera language remarkably well.

Set the lighting and mood

Phrases like 'golden hour', 'soft window light', 'neon-lit street at night' or 'overcast diffuse light' anchor the color palette and atmosphere of the whole clip.

Name a visual style

Add 'cinematic', '35mm film', 'documentary style', 'anime' or 'claymation' at the end of your prompt to lock in a consistent look instead of a generic default.

One scene, one action

Clips are 6–10 seconds. Prompt a single continuous moment — 'a barista pouring latte art' — rather than a sequence of events that can't fit in one shot.

Iterate in Fast, finish in Pro

Draft your prompt in Fast mode at 1.0x credits until the motion looks right, then re-run the winning prompt in Pro or Ultra for the final render. It's the cheapest path to great output.

LTX-2 tutorial FAQ

Keep learning

Ready to make your first LTX-2 video?

Free credits on sign-up, results in minutes, commercially usable output. Compare plans if you need more volume.

Start generating free View pricing