LTX 2.3 · Extend Video

ABOUT THE WORKFLOW

Extend a Video
Upload an existing video and write a prompt describing the next scene. LTX 2.3 reads the final frames, matches the motion, lighting, characters, and style, then generates a seamless continuation with synchronized audio. Use it to add 5 more seconds to a clip that ended too soon, or chain multiple extensions to build a longer sequence. The output is an MP4 with audio that picks up exactly where your original left off.

Partner node. This workflow calls an external API, so each run uses credits from your API wallet. No API key needed. Floyo handles the connection.

Model

• LTX Video 2.3 by Lightricks. An audio-visual video model with native video extension support. Reads the context of your existing clip and generates a seamless continuation with matched motion, character identity, lighting, and synchronized audio.

HOW IT WORKS

Step 1. Upload your video
The clip you want to extend. Any MP4 works: AI-generated, screen-recorded, filmed, or edited.
Works great with: AI-generated clips · short films · product videos · social content · talking-head footage

Step 2. Write the continuation prompt
Describe what happens next. Start with "Continue seamlessly from the previous clip" and then describe the new action, camera movement, and any dialogue. "She turns away from the camera and strolls through the store, the camera follows beside her before moving ahead and turning back to capture her smiling" gives the model a clear direction for the extension.

Step 3. Set the duration
Choose how many seconds to add. The default is 5 seconds per extension.

Step 4. Hit run and download
LTX 2.3 generates the continuation with audio and returns a combined MP4.
Ready for: Premiere · DaVinci Resolve · After Effects · TikTok · YouTube

First time? Upload your video, edit the prompt to match your scene, and hit run. Leave the duration and mode as-is.

RECOMMENDED SETTINGS

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

• Standard video extension (most people) — Upload your video, write a continuation prompt, duration 5 seconds, mode "end." No other changes needed.

• Build a longer video through chaining — Run the workflow once, then feed the extended output back in as the new input. Write a new prompt for the next segment. Repeat to build a sequence of any length, 5 seconds at a time.

• Keep the same action going — Use a prompt that mirrors what is already happening: "Continue seamlessly from the previous clip. The same motion continues, same camera angle, same lighting." This produces the smoothest possible extension.

• Introduce a new action or camera move — Describe the shift clearly: "The camera cuts to a wide shot as she steps outside into the sunlight." The model handles transitions between the existing content and the new direction.

• Character drifts between the original and the extension — Repeat the character description from your original prompt in the continuation prompt. "Same woman, same outfit, same hair" reinforces identity.

• Audio does not match — Describe the sound environment in the prompt: "Same ambient store noise, soft music, her footsteps on wooden floor." The model generates audio to match.

Prompt: Always open with "Continue seamlessly from the previous clip" to anchor the extension to the existing footage. Then describe the new action, camera movement, and audio. Be specific about what stays the same and what changes. The more your continuation prompt echoes the style language of the original (cinematic, photorealistic, warm lighting), the smoother the transition.

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

🎬 Lengthening AI-Generated Clips
Generated a 5-second clip that needs to be longer? Feed it back in and extend it. Chain multiple runs to build 15, 30, or 60-second sequences from a single starting clip.

📱 Social Content Extension
A TikTok or Reel came out too short. Upload it and add 5 more seconds of matching action, camera movement, and audio without reshooting.

🎤 Dialogue and Scene Continuation
Extend a talking-head or dialogue scene. The model picks up the lip-sync, character identity, and ambient sound from the original clip and continues the performance.

🛍️ Product Video Extension
A product reveal ended before the full rotation or hero shot. Extend the clip to complete the sequence with matched lighting and camera motion.

WHAT WORKS BEST / WHAT TO AVOID

✅ Works great

• Clean, stable source videos with clear subjects

• Continuation prompts that echo the style and content of the original

• Gradual action changes (same scene, evolving motion)

• Chaining multiple 5-second extensions for longer sequences

⚠️ May produce softer results

• Source videos with heavy compression artifacts or very low resolution

• Abrupt, unrelated scene changes with no transition language

• Dropping all style keywords from the continuation prompt

• Very fast or chaotic motion at the end of the source clip

FAQ

What does "extend video" mean?
Video extension generates new footage that continues seamlessly from the end of an existing clip. The model reads the last frames, matches the motion, characters, lighting, and audio, and generates a continuation. The result is a longer video that looks and sounds like one unbroken take.

How long can I extend a video?
Each run adds 5 seconds. To go longer, feed the extended output back into the workflow as the new input and run again with a new continuation prompt. There is no hard limit on how many times you can chain extensions.

Does the extension include audio?
Yes. LTX 2.3 generates synchronized audio as part of the extension. It matches the ambient sound, music, dialogue, and sound effects from the original clip and continues them into the new segment.

Can I extend a video I filmed on my phone?
Yes. Any MP4 works as input: phone footage, screen recordings, AI-generated clips, or edited exports. The model reads the content regardless of how it was created.

What does the mode setting do?
The mode controls where the extension happens. "End" adds new footage after the last frame of your clip. This is the default and most common use case.

Does the model maintain character identity across the extension?
Yes. LTX 2.3 reads character appearance, clothing, and proportions from the source video and preserves them in the generated continuation. Reinforce identity by repeating the character description in the continuation prompt.

How to extend a video with AI online?
You can extend a video with AI online through Floyo. No installation, no setup, no API key to wire up. Open the workflow in your browser, upload your video, describe what happens next, 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?
Upload your video, describe what happens next, and hit run.

→ Launch Workflow, Free

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