You can edit a video at two points — before it's generated, on the brief, and after it's generated, on the rendered video. What you can change is different at each stage.
Before generation
Open the video brief and click Edit.
Voiceover — change the words the speaker says, scene by scene. Scene length is derived from the voiceover rather than set by hand, so the editor shows you how much of the scene the line fills (for example "~2.8s of 4s · 8 words").
Headline — change the text shown on the video.
B-Roll — replace, cut or move the B-roll clips.
Chat to edit — describe the change in plain language and the agent rewrites the brief for you. It reaches every field the panes above cover, one at a time. For example:
"Change the headline in shot 2 to…"
"In shot 3, change the line … to …" — keep it about the same length so it still fits
"Put her in a denim jacket" — wardrobe
"Change the location to a modern stadium, and its lighting to golden hour"
"Change what happens in shot 4 to…" — the action in frame
Changes are staged for you to review before they're saved. You can remove any single change, or reset the brief to the original.
When the brief looks right, click Generate.
After generation
Open the rendered video and use the Edit section.
Headline — change the text on the video.
Caption — turn the burnt-in captions on or off.
B-Roll — replace, cut or move the B-roll clips.
Pronunciation — correct a word that sounds wrong. Mark the moment on the waveform, tap the words that came out wrong, add a note, and Segwise regenerates that audio.
Edit AI Video — spot a problem and regenerate. Pin a timestamp on the video, or just describe the issue in plain text, and the agent produces a corrected take. You get a preview and a verdict on whether it's actually fixed before anything changes — nothing lands in your video until you press Apply.
What's locked after generation
The voiceover script and chat-to-edit are both unavailable once a video exists. To change what's said, edit the brief and generate again.