LTX 2.5 for Image to Video

ABOUT THE WORKFLOW

Animate an Image with Sound Upload a starting image and describe the motion you want. LTX 2.5 generates a video clip with synchronized audio in a single pass. Dialogue, sound effects, and ambient noise come built into the output, not layered on afterward.

Model

• LTX 2.5 by Lightricks. A 22-billion parameter open-weights video model that generates picture and audio together, strong at cinematic motion, face detail, and synced sound.

HOW IT WORKS

Step 1. Upload your first frame The image the video starts from. It sets the subject, the scene, and the framing for every frame that follows. Works great with: portraits · landscapes · product shots · concept art

Step 2. Describe the motion (optional) Write a prompt that says what moves and how. Include camera direction, lighting shifts, and sound cues. Leave it empty to let the model decide the motion from the image alone.

Step 3. Hit run and download LTX 2.5 generates the video at half resolution, upscales it, and refines the result in a second pass. You get one video file with synced audio. Ready for: Premiere · DaVinci Resolve · After Effects · any NLE

First time? Leave every setting as-is. The defaults (5 seconds, 16:9, 0.9 MP, prompt enhance on) are the right starting point for almost everyone.

RECOMMENDED SETTINGS

Quick-start guide. Find the goal that matches yours and copy the settings.

• Standard clip (most people) — 5 seconds, 16:9, 0.9 MP, prompt enhance on, random seed. The right starting point for almost everyone.

• Longer scene — Raise duration up to 20 seconds. Past that, motion can drift. Change one setting at a time: raising both resolution and duration together multiplies generation time.

• Higher resolution for delivery — Increase megapixels toward 2.0 MP. The model supports up to 4K HDR. Expect longer generation times at higher sizes.

• Quick test before committing — Drop megapixels to 0.3 or 0.4 for a fast preview. Lock the seed, then raise resolution once you like the motion.

• Reproduce a specific result — Set the seed to a fixed number. Same seed, same prompt, same settings gives the same output.

• Prompt enhance is overriding your intent — Turn prompt enhance off. It rewrites your text for more detail, which helps most of the time but can add motion or elements you did not ask for.

• Motion feels wrong but the look is right — Rewrite the prompt before changing any other setting. Describe what moves and in what direction. "Camera pushes in slowly, hair moves in wind" is better than "cinematic video."

Prompt: Describe the motion, the camera, and the sound. "She lifts the cup to her lips, slow push in, soft jazz piano, rain on cobblestone" gives the model three clear layers to work with. Vague prompts like "make it cinematic" leave too much to chance.

LEARN

📹 Videos

• Intro to Floyo

• ComfyUI 101 Free Course ft. Sebastian Kamph

• Floyo 101 for Team Collaboration

✨ Quick links

• Help docs

• LoRA Training guide

• Discover more workflows

USE CASES

🎬 Filmmakers and Editors Turn storyboard frames or reference photos into moving previews with synced dialogue and ambient sound. Pre-visualize a scene before shooting it.

🎵 Music and Audio Projects Generate a visual companion for a track from a single album cover or mood image. The model produces its own audio layer, which you can replace in post with your own mix.

🛍️ Product and E-commerce Animate a product shot into a short hero clip for a landing page or ad. Camera motion and environmental sound come included.

🎨 Concept Artists and Game Devs Bring static concept art to life. Test how a character, environment, or prop reads in motion before committing to a full 3D or animation pipeline.

📱 Social and Short-form Content Turn a still image into a 5-to-20-second clip for social posts. No editing software needed for a first draft.

WHAT WORKS BEST / WHAT TO AVOID

✅ Works great

• Clear, well-lit single-subject images

• Prompts that name specific motion and direction

• Sound cues in the prompt (dialogue, effects, ambient)

• Matching the resolution aspect ratio to the input image

⚠️ May produce softer results

• Clips longer than 20 seconds (motion can drift)

• Vague prompts with no motion direction

• Heavy bokeh or shallow depth-of-field source images

• Raising resolution and duration at the same time

FAQ

What is LTX 2.5 and who made it? LTX 2.5 is a 22-billion parameter open-weights video model made by Lightricks (released as the LTX company spinout). It generates video and synchronized audio in a single pass from text, image, or video inputs. It was released in August 2026 as the successor to LTX 2.3.

Does LTX 2.5 generate audio with the video? Yes. LTX 2.5 produces audio and video together in one generation pass. Dialogue, sound effects, and ambient noise are generated alongside the picture, not dubbed on afterward. You can include sound cues directly in your prompt to guide what the model produces.

What resolution and length does LTX 2.5 support? LTX 2.5 supports output up to 4K HDR at up to 24 fps and clips up to about 20 seconds. This workflow defaults to 16:9 at 0.9 MP (roughly 1280x736). You can raise megapixels for higher resolution, but generation time increases. Past 20 seconds, motion consistency can drop.

Is LTX 2.5 open source, and can I use it commercially? LTX 2.5 is open weights under the LTX-2.x Community License. It is free for commercial use by individuals and organizations with annual revenue under $10 million. Companies above that threshold need a paid license from Lightricks. This is not an Apache 2.0 or MIT license, so check the terms before deploying at scale.

How is LTX 2.5 different from Wan 2.2 or Kling for image-to-video? LTX 2.5 generates synchronized audio and video in the same pass, which Wan 2.2 and Kling do not. It also supports native multishot generation and up to 4K HDR output. The tradeoff: it is a 22B model, so generation is heavier and needs more VRAM than smaller models.

Do I need a prompt, or can I run it with an image only? You can run it with an image only. The prompt field is optional and empty by default. The model will infer motion from the image. Adding a prompt gives you control over what moves, how the camera behaves, and what sounds appear.

How to run LTX 2.5 image to video online? You can run LTX 2.5 image to video online through Floyo. No installation, no setup, no API key to wire up. Open the workflow in your browser, upload your image, and hit run. Free to try.

WHY FLOYO?

Floyo is the only platform with team collaboration for ComfyUI in the browser. You run workflows with no install. You share run history, assets, and models across your team. You pay only when you generate. Floyo supports open-source and closed-source models.

A director runs a clip and likes the motion. An editor opens that exact run from shared history and keeps going. No file handoffs. No version confusion.

For studios and enterprise teams, Floyo adds private workspaces, pooled resources, and a team usage dashboard. Other ComfyUI cloud tools run for one person at a time. Floyo runs for the whole team, with transparent per-generation costs.

Ready to try it? Upload your first frame, describe the motion and sound, and run it. The settings are already set.

→ Launch Workflow, Free

Questions? Watch the free course or check the FAQ above.