ltx/ltx-2/pro/image-to-video

Transform text or images into cinematic 4K videos with lifelike motion, audio sync, and smooth temporal consistency for professional storytelling at lower computational cost.

Supported formats: jpg, jpeg, png, webp, gif, avif.
The prompt to generate the video from.
The duration of the generated video in seconds.
The resolution of the generated video.
The aspect ratio of the generated video. Only 16:9 is supported.
The frames per second of the generated video.
Whether to generate audio for the generated video.

Introduction to LTX 2 AI Video Generator

Released on October 23, 2025 by Lightricks Ltd, LTX 2 is an open-source AI video foundation model built to redefine how you transform ideas into cinematic motion. Designed for multimodal creation, LTX 2 powers both text-to-video and image-to-video workflows with seamless audio synchronization, native 4K resolution, and up to 50 fps for lifelike fluidity. You can achieve visual and sound coherence in a single generation pass, choose between Fast, Pro, or Ultra performance modes, and enjoy extended clip durations that maintain temporal consistency. Whether you use it for storyboarding, branded content, or visual effects, its lower computational cost lets you produce professional-grade results on consumer-level hardware. LTX 2 image-to-video gives you an effortless way to turn a static image into a living, high-fidelity video clip. It is built for creators, designers, and studios who need fast creative iteration, unified audio-video generation, and 4K quality without technical overhead. With LTX 2 image-to-video, you generate visually dynamic shots in seconds while keeping precise control over motion, camera logic, and style.

Examples of LTX 2 in Action

Video thumbnail
Loading...
Video thumbnail
Loading...
Video thumbnail
Loading...
Video thumbnail
Loading...
Video thumbnail
Loading...
Video thumbnail
Loading...

What makes LTX 2 stand out

LTX 2 converts a single reference image into coherent motion while preserving geometry, materials, and colorimetry. Built for cinematic delivery, LTX 2 maintains temporal consistency across frames, preventing flicker and drift as the scene evolves. With native 2160p support and controllable FPS, LTX 2 balances crisp detail with smooth motion. LTX 2 aligns synthesized movement and optional audio to the source image, yielding believable camera travel, subject dynamics, and environmental responses. LTX 2 emphasizes structure retention and efficient sampling, enabling professional results at lower computational cost. Key capabilities:

  • Structure preservation across pose, layout, depth, and specular highlights.
  • Temporal stability that suppresses texture jitter and identity drift.
  • 16:9 pipeline at 1080p, 1440p, or 2160p with 25 or 50 fps cadence.
  • Prompt-driven motion for camera moves, subject actions, and environment cues.
  • Optional audio generation aligned to on-screen motion timing.
  • Predictable controls: LTX 2 responds consistently to constraints on what to keep or change.

Prompting guide for LTX 2

Start with a sharp, well-lit image_url that is publicly reachable or a base64 data URI. Use a PNG, JPEG, WebP source. In the prompt, describe intended camera motion, subject behavior, and what to preserve. Specify duration (6, 8, or 10 seconds), resolution (1080p, 1440p, or 2160p), and fps (25 or 50) to match delivery cadence. LTX 2 respects the fixed 16:9 aspect, so plan framing accordingly. Use generate_audio to enable synchronized ambience or leave it off for silent cuts. When critical detail must stay intact, tell LTX 2 exactly what to lock; for movement, guide LTX 2 with concise, verifiable actions rather than style-only phrases. Examples:

  • Preserve composition; gentle dolly-in on the subject; background parallax from city lights; duration 10; 2160p; 50 fps; generate_audio true.
  • Studio product on glossy surface; keep reflections crisp; slow 180-degree orbit; subtle softbox flicker; duration 8; 1440p; 25 fps.
  • Portrait in natural light; do not change face; slight head turn and breathing; camera push-in; duration 6; 1080p.
  • Wildlife still; keep foliage; add light breeze movement; handheld micro jitter; duration 8; 2160p; generate_audio false.
  • Night street scene; maintain signage; rain streaks and puddle ripples; slow truck passes behind; duration 10; 2160p; LTX 2 should retain color balance.
  • Fashion shot; preserve silhouette; subtle hair and fabric motion only; duration 6; 1080p; for LTX 2, avoid reframing. Pro tips:
  • Be explicit about constraints such as preserve face, do not alter background.
  • Use spatial language and timing cues like left of subject or first 3 seconds for the move.
  • Keep prompts concise; avoid conflicting adjectives and redundant verbs.
  • Iterate with small revisions to duration, fps, or motion scope before changing the image.
  • LTX 2 favors concrete motion verbs and clear keep vs change directives for stable outcomes.

Related Playgrounds

Frequently Asked Questions

What is LTX 2 and what can it do with image-to-video generation?

LTX 2 is an open-source AI video foundation model by Lightricks designed for generating videos using both text-to-video and image-to-video workflows. It creates synchronized 4K outputs that combine motion, visuals, and sound into one cohesive video.

How does LTX 2 differ from earlier LTX models for image-to-video projects?

Compared with earlier LTX models, LTX 2 introduces synchronized audio generation, native 4K resolution up to 50 fps, and longer 10–20 second clips. These improvements make image-to-video conversions smoother, faster, and more consistent than before.

Is LTX 2 free to use, and how do credits work for image-to-video creation on Runcomfy?

LTX 2 can be accessed in Runcomfy’s AI playground through a credit system. New users receive free credits at signup, and credits are used to generate outputs such as image-to-video results or text-to-video clips. Additional credits can be purchased if needed.

What type of outputs can I expect from LTX 2 when using it for image-to-video generation?

With LTX 2, users can expect high-fidelity 4K video output featuring realistic motion and sound. The system’s image-to-video function can transform still photos or frames into dynamic clips suitable for creative content, ads, and film previsualization.

Who should use LTX 2 for image-to-video workflows?

LTX 2 is ideal for creators, filmmakers, VFX artists, social media producers, and small studios who want to produce high-quality video from static sources or text prompts. Its image-to-video capabilities make it accessible for both professionals and beginners.

What are the hardware requirements for running LTX 2 in image-to-video mode?

LTX 2 is optimized for consumer-grade GPUs, meaning users can run image-to-video generations efficiently without enterprise-level setups. It also supports multi-GPU inference for more demanding rendering needs.

Does LTX 2 support audio output when generating image-to-video results?

Yes, LTX 2 generates synchronized audio and visuals together, unlike earlier AI tools that required audio stitching. This makes image-to-video outputs more immersive and production-ready straight from the model.

Can LTX 2 image-to-video workflows be integrated with other creative software?

Yes, LTX 2 offers APIs and SDKs allowing seamless integration into professional workflows. Creators can connect their image-to-video processes with editing tools, post-production pipelines, or asset management platforms.

Are there any limitations of LTX 2 in image-to-video creation?

While LTX 2 supports up to about 10–20 seconds of synchronized content, very long videos may require chaining multiple runs. Image-to-video outputs also depend on input quality, GPU configuration, and selected mode (Fast, Pro, or Ultra).