Welcome to the trends hub. Every wave of viral AI video gets the same treatment here: what the format is, which tool it runs on, a step-by-step guide, and the fastest way to make your own. This cycle's wave is the AI hug video — two separate photos animated until the people in them embrace, and the rare trend where the participation hook is someone you actually miss.
New to the site? The homepage is the one-screen summary, and the generator is where every guide below ends up.
The format: feed one or two photos into an image-to-video model, describe the embrace, and the model animates the moment two people come together and hug. A grandmother meets her grandchild again. Someone hugs their own younger self. A long-distance couple closes the gap for five seconds. Nothing is filmed, and the comment section does the distribution for you.
Everything we have on this wave:
The hug format splits into recognicable sub-genres, and each one re-lights the trend on its own schedule:
The hub stays constant: each new embrace variant gets an entry here with a guide, a prompt recipe, and the posting notes that matter for it. When a new model ships with better multi-image reference support — or a new hug variant starts climbing the platforms — this page updates first.
Until then, start with the trend guide, grab a template from the prompt guide, and make the one about someone you miss.
Every time a new hug variant or a materially better model breaks. Guides are published while a variant is still climbing, not after it peaks.
The classic two-photo embrace — it is the template every other variant builds on, and the prompt guide gets you from two photos to a finished clip in one sitting.
We track what people actually search for — the question phrasing, the misnomers, the "does this work with one photo" confusion — and build each guide around the real search language.