Edit video with an AI prompt.

Start with the change and a plain verb, then point to the thing so only one thing in the shot matches, like “the guy in the red jacket behind them.” Add a second sentence for what must stay. One line is often enough, and the request box holds up to 2,000 characters. For one real edit followed start to finish, see our step-by-step guide.

You pay in credits: 8 for every second of video at 720p, or 16 at 1080p. The editor tells you the total before you generate. Credits come as one-time packs or weekly plans. New on the web? Aisma can prepare your first three edits before you buy anything: it studies the clip and asks a couple of questions, and the render itself uses credits. See the full price list.

The ones where you can name the target in a few words. Our floating-mug clip is a good model: a steady 10-second shot of a desk, with the mug clear of the notebook and keys around it. A practical tip: trim to the moment you need before you upload. Credits are counted by the second, and a short clip with one clear subject is easier to describe.

It can, for example a new jacket and a snowy street in the same request. One clear change per run is the most predictable, so run the change that matters most on its own, then start the next edit from the finished clip in your Library.

Attach up to 10 photos and call them by their labels, Photo 1 or Photo 2. In the editor, tap a photo’s label to drop it into your request, as in “swap the bottle for the one in Photo 1.” With a single photo, “from the photo” works too. Our product swap examples were each made with one photo, named in the request.

Yes. Put the time in the request, for example “from 0:05 to 0:12, make it snow.” Leave the time out and the change applies across the whole clip.

Fix just the part that missed. On the result page, open Edit a clip, pick that part and add a short note like “keep his T-shirt blue”; its price shows before it runs. To rethink the whole edit, recreate the task with a new prompt. Nothing is charged until you press Generate.

On the web, yes, when it matters. If your prompt leaves a real choice open, such as which photo goes with which person, Aisma asks a quick question before rendering. If nobody replies in about five minutes, Aisma goes with the most likely reading of your prompt.

Yes. The editor says it right next to the request box: write in any language. English works well, and so do Spanish and other languages. Buttons and menus in the app are in English.

Aisma edits the picture in a video you already shot and keeps your original audio, so voices and music play as recorded. Generating clips from text and editing voices or music are both on our roadmap; if you’d use them, email support@aisma.ai with what you have in mind. Until then, a quick phone clip makes a fine starting point.

Yes. The neon city, the cable and the floating mug are real Aisma edits of the clips shown with them, made from words alone, and Full Aisma edit plays each one whole. Examples marked Previous version were recorded with a previous version of Aisma. Before → After views show the original and the edit at the same moment of the clip, never cropped, sped up or retouched. Your result depends on your clip and request.