Most video editing tools treat a clip as a whole. LTX 2 Retake lets you isolate any section, from a single second to several, and describe what you want to change using plain text. The rest of your video stays exactly as it was. You can replace just the visuals, just the audio, or both at once. Tell the model to change the camera angle from a close-up to a wide shot, swap a run for a walk, or add narration over a specific passage. The model applies your instruction only inside the window you define, leaving everything before and after untouched. For video creators working on short films, social content, or product demos, this removes the need to reshoot or record again. Open the clip on Picasso IA, set the start time and duration, write your prompt, and get the revised segment back in seconds.
LTX 2 Retake lets you edit specific moments inside an existing video without reshooting the whole thing. Got a clip where the camera angle felt wrong, the action didn't land, or the pacing needs a fix? You describe the change in plain text, mark the section to rework, and the model generates a replacement that fits into the surrounding footage. On Picasso IA, the whole process runs in your browser with no software to install. It's built for anyone who needs to adjust a recorded video after the fact, from content creators fixing a take to directors reworking a scene.
Do I need programming skills or technical knowledge to use this? No, just open LTX 2 Retake on Picasso IA, adjust the settings you want, and hit generate.
Is it free to try? Yes, you can run LTX 2 Retake directly in your browser without any local setup. Check your account plan for the number of generations included.
How long does it take to get results? Most short-segment edits finish within a minute or two. Longer sections or heavier loads may take slightly more time, but the model runs entirely in the cloud so your device isn't doing the work.
Can I edit multiple sections of the same video? Yes. Run the model once per section, adjusting the start time and duration each time. Reassemble the revised clips in any standard video editor once you have all the parts.
What output formats are supported? The model returns a standard video file you can drop into any editing timeline, upload to a platform, or share directly.
What if the edited section doesn't blend well with the original footage? Refine your prompt to be more specific about the lighting, motion, or mood you want. Even small changes in wording often produce noticeably different results, so iterating quickly is a practical way to get the blend right.
Where can I use the outputs? You can use the clips you generate in social content, client videos, short films, or any other production. Check the terms on your Picasso IA plan for commercial use details.
Everything this model can do for you
Target any timestamp range and apply changes only inside that window.
Describe the edit in plain language and the model interprets your intent.
Choose to replace visuals, audio, or both in a single run.
Everything outside the edited window is preserved exactly as in the original.
Set the edited segment length in seconds to match your pacing needs.
Define where the edit begins so changes land exactly where you need them.
Generate the edited segment directly in the browser without installing any editor.
The train passes the frame, a futuristic city is revealed behind it
replace the shot with a shot on to the runners face
Instead of saying "AI avatars are everywhere, I cant wait to see what will happen next year" make her say "AI agents are everywhere, I cant wait to see them bring my ideas to life"
remove the lens flare
