The two-person AI video — the "duo" format — is the most shared style of AI video on short-form platforms right now. Instead of one character performing to the camera, two subjects trade verses, react to each other, and bounce energy back and forth. It doubles the faces in the thumbnail, doubles the people who share the result, and doubles the comment sections tagging their friends.
The best part: every variant of this format can be made from plain photos in about a minute with ViralReel. Here are the seven formats driving the trend right now, which photos to use for each, and tips to get the best result.
1. The Hotel Lobby AI template (the original)
The format that started the wave: two performers trading verses in the iconic orange booth with the hanging mic. It works for any pair — couples, best friends, siblings, coworkers.
Best photos: one front-facing shot per person, similar lighting on both. Read the full tutorial: How to make the Hotel Lobby AI video.
2. Pet duets
Two pets "performing" the track — cats, dogs, or a cat-and-dog pair — is the most shared sub-format of the trend. Pet content already over-performs on every platform; give it a beat and a recognizable stage and it spreads itself.
Best photos: a level, eye-height photo of each pet. Get down on the floor for the shot. Make sure the face fills a good part of the frame and the fur is well lit.
Pro tip: the pet-owner edition — you and your animal side by side in the booth — is the version that gets "wait, HOW" comments.
3. Baby and family editions
Parents turned the format into a family affair: baby and parent, two siblings, or grandparents and grandkids. These versions consistently pull high watch-through because viewers rewatch to catch each performer's expressions.
Best photos: baby photos are the tricky ones — shoot after a nap, natural window light, no flash, face front and center.
4. Anime edition
For creators with anime profile pictures or fan accounts, the anime edition converts photos into stylized characters performing the same duet. It is the format of choice for fan communities and cosplay crossover content.
Best photos: clear character art or clean selfies that match the character's angles. Cosplay photos also work well.
5. Star edition
The stage-performance variant: same duet energy, more spotlight, more performance posture. Musicians use it to tease covers, and creators use it when they want the "concert moment" look rather than the intimate booth look.
Best photos: front-facing portraits, expressive faces help — the model carries expression into the render.
6. Best-friend reaction duets
The comedic branch of the trend: two friends visibly reacting to each other's lines — the look, the laugh, the side-eye. It is less about singing skill and more about chemistry between the two subjects.
Best photos: pick photos where each person's expression matches their "role" in the bit — one deadpan, one dramatic.
7. Meme and crossover editions
The format is a template in the truest sense: swap in whatever pair makes the joke work — two rival sports logos as images, a "me vs. me" edition with two photos of the same person, historical figures, brand mascot pairs. The most viral crossovers reach audiences far outside the AI niche.
Best photos: whatever makes the pairing funny — but keep both subjects front-facing and well lit, or the render quality falls apart.
Which format should you start with?
| Your photos | Start with |
|---|---|
| You + a friend | Classic booth or best-friend duet |
| You + your pet | Pet-owner edition |
| Two pets | Pet duet |
| A baby in the family | Family edition |
| Anime PFP or cosplay | Anime edition |
| Music or performance content | Star edition |
General tips that apply to every format
- 9:16 vertical, always — this is a short-form native format. Set the ratio in the generator before rendering.
- 8 seconds is usually enough — the trend favors tight loops over long performances. Render 8s first, and go longer only if the bit needs it.
- Post natively per platform — upload directly to TikTok, Reels, and Shorts; cross-posters add watermarks that suppress reach.
- Batch your renders — make three to five variants in one sitting and schedule them. Consecutive posting compounds reach far more than one perfect video.
- Match the sound on-platform — add the official trending audio in the app so the algorithm groups your video with the format.
How much does it cost?
Generation runs on credits, and the cost scales with resolution (480p / 720p / 1080p) and duration (8s / 12s / 15s). The generator always shows the exact credit cost before you confirm, and failed or canceled jobs are refunded automatically. Credits come with every subscription plan, and you can top up any time.
Pick a template, upload two photos, and make your first duo video in about a minute.
