Launch week โ€” 50% off Pro for the first 1,000 creatorsClaim offer
Tutorialsยท4 MIN READ

How to Make an AI Hug Video in 30 Seconds

Turn two portrait photos into a heartfelt hug video โ€” the viral AI effect that's flooding TikTok and Reels. Step-by-step guide.

How to Make an AI Hug Video in 30 Seconds
Photo by Candice Picard on Unsplash

The AI Hug effect is one of the most-shared video templates on TikTok and Reels โ€” and for good reason. Two portrait photos, one prompt, and you get a heartfelt embrace between any two people: reunited family members, friends, fictional characters, anyone.

This guide walks through making one in under 30 seconds, what photo specs work best, and which AI video model handles the effect most reliably.

What you'll need

  • Two portrait photos of the people you want to bring together
  • A Skyvid account (free tier works โ€” no credit card)
  • About 6 credits per generated clip

That's it. No editing software, no prompt engineering experience, no Discord bots.

Step-by-step: 30 seconds to a finished clip

  1. Open the AI Hug effect on Skyvid
  2. Upload your two portraits
  3. Pick clip length (5 or 10 seconds)
  4. Click Generate
  5. Wait ~30 seconds for the model to render
  6. Download โ€” ready to post

The AI handles posing, body language, and the moment of contact. You don't have to write a prompt; the template is pre-tuned.

Photo specs that get the best results

The biggest factor in AI Hug quality isn't the model โ€” it's the input photos. Here's what works:

Resolution and framing

  • Use at least 720ร—720 portraits. Smaller photos produce blurry composites.
  • Front-facing or 3/4 view works best. Side profiles confuse the model.
  • Crop to head + shoulders. Full body shots dilute facial features.

Lighting

  • Similar lighting on both photos. Mixing harsh daylight with studio softbox produces uncanny composites.
  • Avoid heavy shadows across the face โ€” they bake into the output and look wrong when the head moves.

Background

The model auto-replaces the background, so don't worry about matching backgrounds. But:

  • Avoid photos where the background is mistaken for clothing
  • Solid, uncluttered backgrounds give the model less to work around

Which model to use

Kling 2.5 is our default for AI Hug โ€” it nails the body language and facial expression at the lowest credit cost (6 credits per 5-second clip). It's also fast: ~30 seconds end-to-end.

For more cinematic results (slower motion, dramatic lighting), Veo 3.1 is the upgrade. Costs more (12 credits) and takes longer, but the output reads more "film" than "phone."

Hailuo is the alternative for expressive emotional content โ€” slightly different aesthetic, equally strong on character animation.

Prompt customization (optional)

If you want more control beyond the default template, you can add a custom prompt. Useful additions:

  • Setting: "in a park at sunset," "on a busy street," "in a warm kitchen"
  • Mood: "joyful reunion," "tearful embrace," "warm hug"
  • Camera: "wide shot," "close-up on faces," "slow push-in"

Keep prompts under 25 words. Long prompts confuse the template.

Use cases that work

  • Reunion content: long-distance partners, family separated by distance
  • Memorial videos: bringing together photos of loved ones
  • Fan edits: characters or celebrities meeting
  • Storytelling: emotional beats in narrative content

Use cases to avoid

We block AI Hug generation on:

  • Photos of public figures without consent
  • Photos of minors used inappropriately
  • Any content flagged by our safety classifier

This applies to every effect on Skyvid โ€” identity-respecting safeguards are non-negotiable.

Going viral with AI Hug

A few production tips that lift engagement when posting to socials:

  1. Caption with context: "POV: my dad and I after 5 years apart" beats "AI hug video"
  2. Sound matters: pair with a trending emotional audio track
  3. Post at 720ร—1280 (9:16): AI Hug is portrait-oriented by default; don't resize horizontally
  4. First frame counts: pick the thumbnail frame that shows clearest faces

Try it now

Generate your first AI Hug video โ€” free tier includes daily credits, no card required. Most users get a usable clip on the first try.

FAQ

Will the faces look like the original photos? Yes โ€” the model preserves both faces. Front-facing portraits with clear features give the best resemblance.

How long does generation take? About 30 seconds on Kling 2.5. Up to 2 minutes on Veo 3.1 for cinematic results.

Can I use the video commercially? Yes, on paid tiers. Free tier outputs are personal-use only and watermarked.

What if the result looks off? Try different source photos (better lighting, front-facing). Or switch models โ€” Hailuo and Veo 3.1 produce noticeably different aesthetics.

Ready to generate your own?

Free tier ships 10 credits a day โ€” no card required.

Start free
How to Make an AI Hug Video (Step-by-Step Guide 2026) | SkyVid