video
LTX Video
Lightweight LTX image-to-video turning stills into 5-second clips with tunable steps and guidance.
Studio is in early access; availability resolves per project after sign-in.
- Endpoints
- 2
- Eligible
- 2
- Input schemas imported
- 2 / 2
- Delivery
- fal.ai
- Weights
- Not stated
- Developer
- Not stated in source
Overview
LTX Video transforms stills into 5-second clips through image-to-video generation. Thirty inference steps over a 1-50 range balance quality against speed, with guidance scale 2-10 defaulting at 3 for adherence control. Creators get quick extensions from imagery without frame-by-frame animation work.
Capabilities
- 5-second image-to-video clips
- 30-step tunable inference
- Guidance scale 2-10 control
- Base plus dedicated i2v endpoints
Best for
- Social media content
- Product demo animation
- Storyboard prototyping
Use cases
- Rapid social iterations
- Still-to-motion demos
- Extension-free quick turns
Endpoints
2 eligible. Cataloged is not the same as executable: Studio resolves which endpoints a project can run.
| Endpoint | Task | Catalog status | Inputs |
|---|---|---|---|
fal-ai/ltx-video | video editing | Eligible | 5 fields · requires prompt |
fal-ai/ltx-video/image-to-video | image to video | Eligible | 6 fields · requires prompt, image_url |
Questions
Clip length?
5 seconds standard for rapid iteration.
Quality dials?
Steps 1-50 plus guidance 2-10.