How to Make an AI Hugging Video from a Photo (VdoBloom Hugging Video Tab)
Step-by-step guide to VdoBloom's Hugging Video effect: one photo or two, the Warm/Emotional/Gentle/Joyful templates, real credit costs from 10 credits, and how to prompt it.
To make an AI hugging video on VdoBloom, open the Hugging Video tab, upload one photo (AI adds the partner) or two photos of two different people, choose one of the four templates — Warm, Emotional, Gentle or Joyful — and generate; the default Runway Gen-3 at 720p for 5 seconds costs 10 credits, which is exactly the free balance every new account gets. The Hugging Video tab is one of VdoBloom’s keyframe effects, which means your photo is not simply panned and zoomed — it is routed through a keyframe step that stages the embrace before the video model animates it. That is why the hug reads as two people actually closing the distance rather than two flat cutouts drifting together.
What the hugging effect honestly does
Be clear about what you are buying. The effect generates a short cinematic clip in which the person in your photo shares an embrace. It does not resurrect anyone, it does not produce footage of an event that happened, and it is not a memory. It is a generated animation seeded by a still image. The identity likeness is usually strong on a clean frontal photo and gets weaker as the face gets smaller, blurrier or more side-on.
Because these clips animate a photograph of a real person, you must have that person’s consent before you upload their picture — that applies to partners, friends, colleagues and anyone whose photo you did not take of yourself.
One photo or two
The Hugging Video tab has a photo-count toggle, and the two modes behave differently:
- One photo. You upload a single person and the model invents the partner. The prompt VdoBloom sends says the person from the uploaded image is hugging an AI-generated partner, with the explicit instruction that only these two people appear in the scene. Use this when you want a warm, generic embrace and you do not care who the other person is.
- Two photos. You upload Person 1 Photo and Person 2 Photo, and the prompt becomes the person from the first image and the person from the second image hugging each other. This is the mode most people actually want, and it is the one that needs both subjects’ consent.
In two-photo mode the tab keeps the native effect roster — Runway, Kling, PixVerse and Vidu — because those are the model families whose endpoint accepts a second image. Do not expect the whole 105-model catalogue here; the picker deliberately narrows to what can genuinely consume both photos.
What photo works best
The single biggest quality lever is the source image, not the model. What works:
- Head and upper body in frame, roughly waist-up. A full-length shot shrinks the face; a tight head crop leaves the model no arms to animate.
- Both arms visible and unobstructed. Arms folded, hands in pockets or a bag across the chest all fight the embrace.
- Even, soft light on the face. Harsh backlight and heavy shadow are the usual cause of a melted-looking face mid-motion.
- Neutral or slightly smiling expression. An extreme expression gets baked into every frame.
- Matching framing between the two photos in two-photo mode. If one subject is a close-up and the other is full-length, the composite scale looks wrong.
The four templates
| Template | Mood | Lighting it asks for | Best for |
|---|---|---|---|
| Warm | Comforting, affectionate | Soft lighting | The safe default; couples, family, anniversary posts |
| Emotional | Deep, heartfelt | Dramatic lighting | Reunion and tribute edits where you want weight |
| Gentle | Tender, caring | Warm lighting | Parent and child, quiet comfort, slower cuts |
| Joyful | Happy, celebratory | Bright lighting | Friends, graduations, welcome-home clips |
Each template writes a full prompt for you, so you never have to start from a blank box. The editable default prompt behind the tab is simply: warm hugging scene, soft lighting, affectionate moment, cinematic. Picking a template swaps in a longer version of that with the people-count constraint attached.
What it costs in credits
There is no surcharge for the keyframe step. You pay the standard price of whichever video model you pick, and one credit is $0.05 on the Lite plan ($15 for 300 credits). These are the real numbers for the models on this tab:
| Model | Tier | Credits | Cost at $0.05/credit | Notes |
|---|---|---|---|---|
| Runway Gen-3 | 720p, 5s | 10 | $0.50 | Tab default; fits the 10 free signup credits exactly |
| Runway Gen-3 | 1080p, 5s | 20 | $1.00 | Best value upgrade |
| Runway Gen-3 | 720p, 10s | 20 | $1.00 | Double the length for the same price as 1080p |
| Runway Gen-3 | 1080p, 10s | 30 | $1.50 | Full-length, full-resolution |
| Vidu Q2 Turbo | 720p, 4s | 15 | $0.75 | Cheapest non-Runway option here |
| Vidu Q2 Turbo | 1080p, 5s | 22 | $1.10 | Supports background music and movement amplitude |
| PixVerse v5 | 720p, 5s | 25 | $1.25 | Style presets; 5s and 8s durations |
| PixVerse v5 | 1080p, 8s | 50 | $2.50 | Longest PixVerse tier on this tab |
| Kling V2.5 Turbo | 5s | 40 | $2.00 | No quality toggle — one tier, strong motion |
| Kling V2.5 Turbo | 10s | 80 | $4.00 | Priciest option; save it for a hero shot |
Credits only leave your balance after the provider accepts the job, so a rejected request does not silently bill you. One-time credit packs start at $2.49 for 75 credits and never expire, and annual billing brings Lite down to $10 a month. Full plan detail sits on the pricing page.
The exact steps
- Go to the Hugging Video tab.
- Set the photo count: one photo for an AI-generated partner, two for two real people.
- Upload. In two-photo mode the slots are labelled Person 1 Photo and Person 2 Photo — the order matters only for how the prompt refers to them, not for who leads the hug.
- Pick a template. Warm first if you are unsure; the preview clip on each card shows the motion you will get.
- Choose your model and tier. Start at Runway Gen-3 720p 5s for the cheapest honest test.
- Set the aspect ratio — 9:16 for Reels, TikTok and Shorts, 16:9 for YouTube, 1:1 for a feed post.
- Edit the prompt only if you want a specific setting (see below), then generate.
- Download. Free accounts get a watermark; any paid plan downloads clean with commercial rights.
Prompting it well
The templates already carry the hard constraints, so your edits should add setting and camera, not re-describe the hug. Add a location (airport arrivals hall, kitchen at golden hour, hospital corridor), a camera instruction (slow push-in, static wide, gentle orbit) and a wardrobe or season cue if it matters. Leave the phrase about only these two people in the scene alone — removing it is the most common cause of a stray third person appearing in the frame.
Avoid stacking too many actions. A hug plus a spin plus a kiss plus a camera whip in five seconds gives the model no time, and the result is a smear. One motion per clip, then cut.
When to pick a different effect
If the moment you want is romantic rather than affectionate, the Kissing Video tab uses a stricter identity-preservation prompt and is the better fit. If you want a spiritual comfort clip, the Jesus Hug tab is purpose-built for that and takes one photo. And if you only want the person in your photo to move — a wave, a laugh, a turn — a plain image-to-video run is cheaper than any keyframe effect. You can test all of this against the free tier before spending anything, and Vidu Q2 Turbo is the cheapest paid rung if 10 free credits are not enough.
Frequently asked questions
Can I make a hugging video of someone who has passed away?
Technically yes, from a clear photo, and this is one of the most common uses of the effect. Treat it as a piece of generated art rather than footage, tell your audience it is AI, and make sure the family is comfortable with it before posting.
Does the free plan cover a hugging video?
Yes. A new account gets 10 free credits with no card, and Runway Gen-3 at 720p for 5 seconds costs exactly 10 credits. That is one complete hugging video, with a watermark on the download.
Why does the second person look different from my photo?
Identity drift is usually a source-image problem: small face, motion blur, strong side angle or heavy shadow. Re-crop to waist-up, use the sharpest frontal photo you have, and try 1080p — more pixels give the keyframe step more to hold on to.
Can I get a longer hugging video?
The tab tops out at 10 seconds on Runway Gen-3 and Kling V2.5 Turbo, and 8 seconds on PixVerse v5 and Vidu Q2 Turbo. For anything longer, generate two clips with different templates and cut them together.
Is the output watermark-free?
On any paid plan, yes, and it comes with commercial rights. Free-tier downloads carry a VdoBloom watermark.
Ready to try it?
Create your first AI video in minutes — no credit card required.
Start Creating Free →