LTX Director 2 · Timeline Video With Audio
Direct a multi-shot video on a timeline with LTX Director 2 and LTX-2.3. Drop images at the points you want, write what happens between them, hit run.
image to video
ltx 2.3
ltx director
ltx director 2
timeline video
53
ABOUT THE WORKFLOW
Direct a Video on a Timeline Place your images on the timeline at the points the video has to pass through. Write what happens between them. The model draws picture and sound together in a first pass, upscales and refines in a second, and saves one clip with audio as a single file.
Model
LTX-2.3 22B distilled by Lightricks. The same open-weight model that powers the LTX-2.3 image to video workflow, running through a two-stage pipeline: a fast 8-step draft pass, then a 4-step detail pass with a spatial x2 upscaler. Audio and video are written in the same pass.
LTX Director 2 timeline node by WhatDreamsCost. A community-built timeline editor that replaces the usual node graph with an interface closer to Premiere or After Effects, where images, prompts, and audio sit on tracks and can be trimmed and moved.
HOW IT WORKS
Step 1. Replace the images on the timeline Two reference images sit on the main track already. Swap them for your own. Where you place one is where the video reaches it. Works great with: character art · storyboard frames · product shots · location photos
Step 2. Write the global prompt Set the look across the whole clip. Setting, lighting, film style, and what must stay consistent. This is the mood board, not the shot list.
Step 3. Write a prompt for each segment Click a segment on the timeline and describe what happens during that stretch. "She turns to the window, slow push in, curtains drift." Each segment gets its own instruction.
Step 4. Add audio or a motion reference (optional) Both tracks are on and empty. Drop in a voice clip, a music file, or a motion reference video and the model works from it alongside your images and prompts.
Step 5. Hit run and download A draft pass builds the clip, then it is upscaled and refined in a second pass. The result is saved under video/LTX_Director as one file with sound. Ready for: Premiere · DaVinci Resolve · CapCut · After Effects
First time? Leave every setting as-is. The defaults (5 seconds · 24 fps · 8+4 steps · fixed seed) 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 · 24 fps · 8+4 steps · seed 0 fixed. The right starting point for almost everyone.
Longer clip — Drag the segment blocks further along the timeline. Duration is set by how far the blocks run, not by a number you type.
More shots in one clip — Add more images to the main track. Each one becomes a point the video has to pass through, and the model builds the motion between them.
Want your own audio on the track — Drop a voice clip or music file onto the audio track. The model writes picture and sound around what you give it rather than inventing its own.
The motion between shots is wrong — Rewrite the segment prompt for that stretch. Say what moves and how the camera behaves, not what is in the frame.
The look drifts across shots — Tighten the global prompt. Name the setting, the lighting, and the character design that must hold from start to finish.
Repeat a take you liked — The seed ships at 0 fixed. Change the number when you want a different take, or switch to random for variety.
An image's prompt box will not open — An image needs at least one second of space on the timeline before its segment prompt appears. Stretch the block.
Guidance does not seem to help — Leave it at 1. The distilled model is tuned for that value and raising it flattens the output.
Prompt: Split the job in two. Global prompt sets the look: "cinematic anime rooftop, golden hour, warm rim lighting, consistent character design throughout." Segment prompts set the action: "he crouches and frames the camera with his hands, slow push in, rooftop wind." Keeping look and action separate gives you control over each without rewriting the other.
LEARN
📹 Videos
ComfyUI 101 Free Course ft. Sebastian Kamph
Floyo 101 for Team Collaboration
✨ Quick links
USE CASES
🎬 Short Films and Trailers Lay out a sequence of shots on the timeline with transitions and audio, and get one continuous clip back instead of stitching separate renders.
🎵 Music Videos Drop a music track on the audio line, place keyframe images at the beats, and let the model fill in the motion between them.
📖 Storyboard to Motion Turn a set of storyboard frames into an animatic with camera moves and room tone, without leaving ComfyUI.
🛍️ Product and Brand Films Build a multi-shot product spot from still renders, controlling which frame appears at which second.
WHAT WORKS BEST / WHAT TO AVOID
✅ Works great
A small number of clear reference images placed at key beats
Global prompts that pin the look and segment prompts that describe action
Steady, readable camera moves named in each segment
Audio tracks that the model can sync motion to
⚠️ May produce softer results
Many tightly packed images with no room for the model to build motion
Segment prompts that describe the frame instead of what happens
Fast cuts or story jumps that ask for logic the model cannot follow
Readable text inside the frame, where LTX-2.3 is weaker
FAQ
What is LTX Director 2? LTX Director 2 is an open-source ComfyUI timeline node built by WhatDreamsCost that turns LTX-2.3 video generation into a visual editing workflow. Instead of wiring separate loader, sampler, and audio nodes by hand, you place images, text prompts, and audio on tracks and trim them the way you would in Premiere or After Effects. The model reads the whole timeline as one clip and generates picture and sound together.
What model does this workflow use? LTX-2.3 22B distilled by Lightricks, the same open-weight model behind the standalone LTX-2.3 image to video workflow. It runs a fast 8-step draft pass, then upscales with a spatial x2 model and refines in a 4-step detail pass. A Gemma 3 12B text encoder reads the prompts, and separate video and audio VAEs decode the result.
Can I add my own music or voice track? Yes. The audio track is on and empty by default. Drop a file onto it and the model writes picture and motion around your audio rather than generating its own. You can also add a motion reference video to a separate track.
How long can a clip be? Duration is set by the timeline, not by a fixed number. This workflow ships at 5 seconds, but stretching the segment blocks extends the clip. LTX-2.3 itself supports up to 20 seconds, though generation time and memory climb with length.
Is LTX Director 2 free for commercial use? The LTX Director node is open source from WhatDreamsCost. The LTX-2.3 model underneath is released by Lightricks and is free for commercial use for organizations under 10 million USD in annual revenue. Above that threshold a commercial license from Lightricks is required. Check the license text before building on it.
How does this compare to the standard LTX-2.3 image to video workflow? The standard workflow takes one start image and one prompt and returns a single continuous shot. LTX Director 2 adds a timeline where you arrange several images at specific points, write a prompt for each stretch, and drop audio on a separate track. Use the standard workflow for a single shot, and this one when you want to direct where the video goes and when.
How to run LTX Director 2 online? You can run LTX Director 2 online through Floyo. No installation, no setup, no model downloads. Open the workflow in your browser, place your images on the timeline, write the prompts, 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 designer runs an edit and likes the result. A teammate 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? Place your images on the timeline, write what happens between them, and run it. The settings are already set.
Questions? Watch the free course or check the FAQ above.
Read more
