Start every LTX-2 video from your own photo
Upload a photo and it becomes the opening frame of your clip. It starts on your product and your lighting, just as you shot them. Describe only what should move, and the model adds the motion.
LTX-2 image to video with sound
Animate a photo with LTX-2 by Lightricks. Describe the motion and get a clip with sound made in the same pass, vertical or wide to match your shot.
Model
LTX-2 on Pixelcut
LTX-2 by Lightricks turns one starting photo and a motion prompt into a 5 to 15 second MP4 at 30 fps, with sound on by default. Switch to Turbo for faster, lower-cost runs. Both run on Pixelcut Pro.
LTX-2 at a glance
- Input
- Photo + promptStart photo required
- Resolution
- Matches your photo
- Aspect ratio
- Matches your photo
- Output
- MP4
- Clip length
- 5 to 15 s
- Audio
- On or offSound generated with the video
LTX-2 Turbo at a glance
- Input
- Photo + promptStart photo required
- Resolution
- Matches your photo
- Aspect ratio
- Matches your photo
- Output
- MP4
- Clip length
- 5 to 15 s
- Audio
- On or offSound generated with the video
Ideas to bring to life with LTX AI
Upload a photo, describe the motion, and LTX-2 turns it into a short clip with sound.
Image to video
Add a photo, then say what moves
LTX-2 on Pixelcut starts from a photo you already have. Upload it, add one line on the movement, the camera and the sound you want, and press generate. Keep that line to one action and one camera move for the cleanest result.
- Upload the photo you want to bring to life
- Describe the motion, the camera move and the sound
- Generate the clip and download the MP4
Write LTX-2 prompts that describe the shot you want
Get closer to the shot you picture by naming the action first, then the camera move, the light and the sound. Short, specific lines give cleaner clips than long lists of adjectives. For sound, name what makes it rather than the mood.
Test ideas fast with LTX-2 Turbo, then pick your take
Turbo is the faster, lower-cost version of the same model, with the same options. Try a few takes of an idea in quick runs and keep the one that works. For a second look, run the same photo on the standard version.
Pick the clip length that fits your post
Set the length one second at a time, from 5 to 15 seconds. Go short for a teaser or longer for a full reveal. Your clip keeps the shape of your photo and drops straight into your edit.
Generate LTX AI video and sound in one pass
Make video and sound together from one prompt. Name the sound you want in the same line as the motion, and the audio comes back in the clip. You get a finished scene ready to share, with no separate sound step.
Who uses the LTX-2 AI video generator
Filmmakers, marketers and social creators use LTX-2 to turn the photos they already have into short clips with sound.
Film and storyboard teams
Bring concept art and storyboard panels to life for a pitch or client review. Ambient sound sells the mood before anything is shot.
Brands and marketers
Turn a clean product photo into a campaign teaser with its sound in place. Show the product in motion before you book a shoot.
Social creators
Animate your travel photos and pet pictures for short-form feeds, with sound already in the clip. Want to compare models? Try the AI video generator.
More models on Pixelcut
Switch models per generation instead of per subscription. Each card opens that model's page.
Cinematic motion, and your first video is free
- New
- Free with sign-up
- Image to video
- Up to 10 s
ByteDance's newest video model
- New
- Audio
- Image to video
- Up to 30 s
Text, image and reference video, up to 15 s
- New
- Audio
- Image to video
- Up to 15 s
Photoreal video with synchronized audio
- Audio
- Image to video
- Up to 8 s
Multi-shot scenes and elements
- New
- Audio
- Image to video
- Up to 15 s
Synchronized audio in many languages
- Image to video
- Up to 10 s
How to use LTX-2 AI to turn a photo into a video
- 1
Upload your starting photo
Pick the image you want to animate, such as a product shot, portrait, travel photo or illustration. Its shape sets the frame of the finished video.
- 2
Describe the motion
Write what should move, how the camera travels and what you want to hear, like a slow push in on a campfire with crackling logs and crickets at night.
- 3
Pick model, length and sound
Choose a duration in one-second steps, keep audio on to hear sound with your clip, and pick Turbo when you want faster, lower-cost runs while you test ideas.
- 4
Generate and download
Generate the clip, play it back with sound and download the MP4. Want a different take? Change a few words in the prompt and run it again.
Plans for video generation
Video is billed in credits per second of footage. Pick the plan that matches how much you make.
Save 20% with a yearly plan
Free
$0
per month, billed yearlyYour first video free on MiniMax H3 Max
Image tools free to try
Pro
$10$8
per month, billed yearly600 AI Credits Monthly
Access all AI models
Unlimited Background Removal
Unlimited Upscale
3 Person Team
1,000 Batch Exports Monthly
Commercial license
Most Popular
Business
$30$24
per month, billed yearly3,600 AI Credits Monthly
Access all AI models
Unlimited Background Removal
Unlimited Upscale
10 Person Team
2,000 Batch Exports Monthly
Commercial license
Trusted by over 200,000 creators, brands and small businesses
Trusted by creatives at
Frequently Asked Questions
LTX-2 is an AI video model from Lightricks that creates picture and sound together. Pixelcut runs the 19B release, plus a faster Turbo version, to turn a starting photo into a short MP4 with audio.
It suits short clips where sound matters, because picture and audio come out of one generation. On Pixelcut it always starts from a photo you upload. Pick it when you already have the shot.
Yes. Sound is generated in the same pass as the picture, and you get one clip with audio instead of adding a track later. Audio is on by default and can be switched off.
On Pixelcut, clips run from 5 to 15 seconds, set in one-second steps. Every video renders at 30 fps and downloads as an MP4, with the aspect ratio taken from your starting photo.
Turbo is a faster, lower-cost version with the same options. Pixelcut's model picker estimates about one minute per Turbo run and about two minutes for the standard version. Turbo suits quick testing.
Pixelcut connects LTX-2 and LTX-2 Turbo to Claude, ChatGPT, Codex and Cursor through the Pixelcut MCP server. Each LTX-2 request sends one start photo to animate with your prompt, and every clip uses Pixelcut credits.
