Nidhogg
Log in

AI Video Editor

Nidhogg's AI video editor works by instruction, not timeline. Upload a clip, type the change — 'make it golden hour', 'swap the hoodie for a tan trench coat', 'remove the crowd in the background' — and Kling O1 Edit re-renders the shot with that change applied while the performance, framing, and camera movement carry through.

It's closest to a VFX pass you can direct in plain English. It won't cut your footage or mix your audio — that's still a timeline's job — but the category of change that used to mean rotoscoping, keying, or a reshoot now takes one sentence and a short wait.

Edit a clip with a prompt
AI Video Editor

What prompt-instructed editing means

Traditional editors manipulate the recording; a prompt-instructed editor regenerates it. The model reads your clip, understands what's in it, and renders a new version with your instruction applied — the same person walks the same path through the frame, but it's raining now, or the storefront sign is gone, or the whole shot has become dusk.

Because the output is a re-render, review it like a new take: check that faces, logos, and details you didn't mention came through faithfully. Most edits do; when one drifts, rerunning with 'keep everything else unchanged' appended to the instruction tightens it up.

Edits it handles well

Environment changes are the sweet spot: time of day, weather, season, and background swaps hold up because the subject's motion anchors the shot. Wardrobe and color changes on a clear subject are similarly reliable — 'change the car to matte black', 'make her jacket red' — as is removing background distractions.

Style transformations work too — 'render this as hand-drawn animation' — with the honest caveat that heavier restyles vary more from run to run. Generate two or three takes of a big transformation, pick the winner, and finish it with the video upscaler.

Where it fits next to a timeline editor

Think of it as the shot-fixing layer under your normal edit. You still assemble, trim, and mix in a timeline tool; Nidhogg handles the changes inside the frame that a timeline can't touch without compositing skills. A practical loop: pull the shots that need fixing, run each through a prompt edit, then drop the fixed versions back into your cut.

The surrounding tools cover the rest of finishing: reframe converts a shot between aspect ratios when you need a vertical from a horizontal master, and the video upscaler brings the final clip up to delivery resolution.

Writing instructions that work

One change per pass. 'Make it night and remove the van and change his shirt' forces the model to juggle; three focused passes are more controllable and easier to judge. Name things the way you'd point them out to a person on set: 'the blue sedan parked on the left', not 'the vehicle'.

State what should stay: 'keep the camera movement and the subject's face unchanged' does real work in an instruction. And keep clips short — Nidhogg processes shots that run seconds, not minutes, so edit per shot and assemble the sequence afterward.

FAQ

Is this a replacement for a timeline editor?+

No, and it isn't trying to be. It doesn't cut, trim, or handle audio. It changes what's inside the frame — the part timeline editors can't touch without VFX work. Most people use both: fix shots here, assemble there.

Can I edit footage I shot myself?+

Yes. Upload any short clip you have the rights to — phone footage, camera footage, or clips generated on Nidhogg.

Does it preserve the original motion?+

That's the point of the tool: the performance and camera path carry into the re-render while your instructed change is applied. Verify fine details like faces and on-screen text after any edit, since the output is a regeneration rather than a patch.

How long can the clip be?+

Edits run on short clips — individual shots of a few seconds, matching how Nidhogg generates video. For a longer piece, edit shot by shot and assemble in your timeline.

Can it remove objects or people?+

Yes — removal is one of the most-used instructions. Describe the object and its location, and the model re-renders the clip without it, rebuilding the background behind where it stood.

Ready to try it?

Free credits on signup — no card required.

Edit a clip with a prompt

Related