How to Make an AI Lift and Carry Video With Two Photos on VdoBloom
How VdoBloom's Lift & Carry effect works: one photo of two people or two separate uploads, the three real templates, which models truly read both photos, and costs from 8 credits.
To make an AI lift and carry video on VdoBloom, open the Lift & Carry tab, upload either one photo containing two people or two separate person photos, choose the Playful, Dramatic or Sweet template, and generate — the cheapest working configuration is Runway Gen-3 at 720p for 5 seconds at 10 credits, and the cheapest Seedance option is Seedance 2 Mini at 480p for 5 seconds at 8 credits. The effect animates one person lifting and carrying the other.
What makes this tab different from the one-photo effects
Most VdoBloom effect tabs take a single photo. Lift & Carry is a two-person effect, so it ships with a photo-count toggle and two upload slots labelled Person 1 Photo and Person 2 Photo. You choose the mode that matches what you have:
- One photo, two people. You already have a shot of both subjects together. The model works from that single frame, which keeps identities and lighting consistent because they were photographed in the same scene.
- Two separate photos. One person per upload. The model composes them into one scene and animates the lift. This is the flexible option, and the harder one to get right.
Lift & Carry is a direct effect, not a keyframe effect — your photos go straight to the chosen video model rather than through the WAN 2.7 keyframe stage that tabs like Laughing and Bikini use. That is why its model roster is wider than those tabs, but also why the second photo matters so much: there is no intermediate keyframe repairing a mismatched pair.
If either photo shows a real person, you need that person’s consent before you animate them. A lift and carry clip puts two identifiable people in physical contact, which makes consent for both subjects the baseline, not a courtesy.
Not every model can honestly use both photos
This is the detail that decides whether your second upload does anything at all. In two-photo mode the picker narrows to models that genuinely consume two images:
- Native effect models — Runway Gen-3, Gen-4 Turbo and Gen-4.5, Kling V2.5 Turbo, the PixVerse v3.5 to v6 line, and the Vidu 2.0 / Q1 / Q2 / Q3 line. These take the second image directly on the effects endpoint.
- The Seedance 2 family — Seedance 2, Seedance 2 Fast, Seedance 2 Mini and Seedance 2.5, which accept both photos as multimodal reference images tagged in the prompt.
- Veo 3 and the Gemini Omni family, which accept multiple reference images natively.
Deliberately excluded: Kling 3.0 Omni, because its second image means end frame — the video would morph person one into person two instead of showing them together. Wan 3.0 is out for the same first-frame/last-frame reason. If you pick a model and the second upload slot stops mattering, that is the reason.
The three templates and their real prompts
Templates are saved prompts with preview clips, and Playful is the tab default. These are the exact strings:
- Playful — lift and carry, playful romantic moment, strong arms, joyful expression, cinematic
- Dramatic — dramatic lift and carry, cinematic moment, powerful arms, intense emotion, dramatic lighting
- Sweet — gentle lift and carry, sweet romantic moment, soft lighting, tender expression, cinematic
Playful gives the widest motion and the brightest grade. Dramatic swaps natural light for a hard key and slows the movement, which hides limb artefacts better than the other two. Sweet is the lowest-amplitude option and the most forgiving of imperfect source photos, because less movement means fewer frames where anatomy can break.
Credit cost by model
One credit is $0.05 at the Lite rate of $15 for 300 credits. These are the real costs for the models available on this tab:
| Model | Tier (5-second clip) | Credits | Cost at $0.05/credit | Best for |
|---|---|---|---|---|
| Seedance 2 Mini | 480p | 8 | $0.40 | Cheapest test of a two-photo pairing |
| Runway Gen-3 | 720p | 10 | $0.50 | Default; best cheap native two-photo path |
| Seedance 2 Mini | 720p | 17 | $0.85 | Budget clip you can actually post |
| Vidu Q2 Turbo | 720p | 18 | $0.90 | Fast renders, smooth bodies |
| Runway Gen-3 | 1080p | 20 | $1.00 | Delivery quality at low cost |
| Vidu Q2 Turbo | 1080p | 22 | $1.10 | Cheapest 1080p after Gen-3 |
| Seedance 2 Fast | 480p | 24 | $1.20 | Seedance prompt control, lower spend |
| PixVerse v5 | 720p | 25 | $1.25 | Punchier motion on the lift |
| Seedance 2 | 480p | 38 | $1.90 | Best identity handling of two photos |
| Kling V2.5 Turbo | no quality tier | 40 | $2.00 | Strongest full-body physics |
| Veo 3 Fast | flat | 40 | $2.00 | Multi-reference with native audio |
| Runway Gen-4.5 | no quality tier | 48 | $2.40 | Most cinematic frame |
| Seedance 2 Fast | 720p | 50 | $2.50 | Sharp Seedance output |
| Seedance 2 | 720p | 82 | $4.10 | Hero clip, two distinct faces held |
| Seedance 2 | 1080p | 203 | $10.15 | Maximum fidelity, spend last |
Kling V2.5 Turbo at 10 seconds is 80 credits, and Veo 3 Quality is 160 credits flat — both worth knowing before you change a dropdown by accident. Credits are only deducted once the provider accepts the job, and a generation that fails after acceptance is refunded automatically. The plan and pack ladder is on the pricing page; packs start at $2.49 for 75 credits and never expire.
Step by step
- Open the Lift & Carry tab.
- Set the photo-count toggle. One photo of two people is the higher-success path; use it when you have it.
- Upload. In two-photo mode, put the person doing the lifting in Person 1 Photo and the person being carried in Person 2 Photo.
- Pick a template. Start with Sweet if the photos are imperfect, Playful if they are clean full-body shots.
- Choose a model. Run the first attempt on Seedance 2 Mini at 480p (8 credits) or Runway Gen-3 at 720p (10 credits).
- Set duration, resolution and aspect ratio. Gen-3 offers 5s and 10s, 720p and 1080p, and 16:9, 4:3, 1:1, 3:4 or 9:16.
- Review, then re-run the winning combination on Seedance 2 or Kling V2.5 Turbo only once you know the pairing works.
Getting a believable lift
Two-person physics is the hardest thing to ask of an image-to-video model, and photo choice does most of the work.
- Show legs. Head-and-shoulders crops give the model nowhere to put a carry. Full-body or knee-up shots produce dramatically better results.
- Match the light. In two-photo mode, an outdoor daylight photo paired with an indoor tungsten photo will composite badly no matter which model you pick.
- Match the scale. Two photos shot at wildly different distances give the model conflicting size cues, and you get a child-sized adult.
- Add camera direction, not emotion. The templates already carry the emotion. Add full body in frame, static camera to stop the model cropping the lift out of shot.
When another tab fits better
Lift & Carry is one of a small group of two-person effects, and picking the right one beats prompt-wrestling this one:
- Bridal Carry for the specific wedding-threshold pose.
- Couple Dance for waltz, salsa or slow-dance motion between two people.
- Carry Me for a piggyback-style carry rather than a front lift.
Before spending on a premium run, read the spec pages for Runway Gen-3 and Kling V2.5 Turbo, the two models most people end up choosing here. If you are delivering the clip to a client, the no-watermark guide covers which plans strip the watermark and grant commercial rights.
Frequently asked questions
How much does an AI lift and carry video cost?
From 8 credits ($0.40) on Seedance 2 Mini at 480p for 5 seconds, or 10 credits ($0.50) on the default Runway Gen-3 at 720p. A premium Seedance 2 run at 720p is 82 credits ($4.10).
Do I need two photos?
No. The tab has a photo-count toggle: one photo containing two people works, and is usually the more reliable route because lighting and scale already match.
Why is my second photo being ignored?
Almost always a model choice. Only the native effect models, the Seedance 2 family, Veo 3 and the Gemini Omni family consume two photos on this tab. Kling 3.0 Omni and Wan 3.0 treat a second image as an end frame, so they are excluded from two-photo mode on purpose.
Which model handles two different faces best?
Seedance 2, because it takes both uploads as tagged reference images rather than a single composite. It costs 38 credits at 480p and 82 at 720p for 5 seconds, so prove the pairing on a cheaper model first.
Can I use photos of people I know?
Only with their consent — both of them, since both appear in the clip. Explicit and illegal content stays blocked on every model and tab.
Ready to try it?
Create your first AI video in minutes — no credit card required.
Start Creating Free →