AI photo and video tools use a handful of terms that are rarely explained. This glossary defines each one in a sentence or two, in the order you are likely to meet them.
Generating a new image from an existing photo. The source photo supplies the person; the model builds a new scene, pose or outfit around them. Every photo result on MiraNight is image-to-image.
Generating a short moving clip from a single still photo. The model predicts the motion frame by frame while keeping the person consistent. Clips are short, usually a few seconds, because consistency gets harder with every extra frame.
How closely the generated person resembles the person in the source photo: face shape, eyes, nose, lips, hair and skin tone. It depends mostly on the input photo; a sharp, front-facing face gives the best likeness.
A ready-made scene — setting, pose, lighting and outfit — that is applied to your photo. You choose a style instead of writing instructions yourself.
The photo you upload. It should show one person with a clearly visible face; on MiraNight it can be a JPG or PNG of up to 20 MB.
The size of a result in pixels. For video it is usually given by its height, such as 480p or 720p. A higher resolution shows more detail and takes more computing time.
A video is a sequence of still images called frames; the frame rate is how many of them are shown per second. Every frame has to be generated, which is why a video takes longer and costs more than a photo.
How long a generation takes. A photo is usually ready in about 25 seconds; a short video takes a few minutes.
When many people create at the same time, jobs wait their turn. On MiraNight you can have up to two jobs waiting at once, and the page shows your position.
The unit you pay with on MiraNight. Every creation shows its credit cost before you confirm; credits never expire and there is no subscription.
If a creation fails or cannot be finished, its credits are returned automatically.
A visible mistake in a generated result, such as an extra finger, a blurred hand or a warped background. Artefacts are less likely with a clear input photo and a simple pose.
Agreement from the person in the photo. On MiraNight you may upload only your own photo or one of an adult aged 18 or over who has explicitly agreed.
Image-to-image produces one new still picture; image-to-video produces a short moving clip. Both start from the same single photo.
Every extra frame is another chance for the face or body to drift away from the original. Short clips keep the person consistent.
It shows more detail, but a result can only be as good as its input photo. A sharp face at 480p often looks better than a blurred one at 720p.
🎁 100 free credits💳 No card needed
Continue with GoogleLog in with your Telegram account — your balance and history connect to the site. Tap the button below, then confirm the login in the chat.
Open Telegram and confirmThis site contains AI-generated content intended for adults only. By continuing you confirm you are over 18 and legally permitted to access this content in your country.