AI Carry Me Video Generator: Piggyback and Fireman Carry Clips
The VdoBloom AI Carry Me Video Generator makes fun carrying videos from photos - piggyback rides and fireman carries, generated rather than filmed. It is the unserious member of the carry family. Where bridal carry aims for romance, this one aims for the reaction you get when you send it to the friend who is suddenly being hauled over someone's shoulder in a video that never happened.
Two templates cover the classics. Piggyback Ride animates one person carrying the other on their back, joyful and slightly chaotic. Fireman Carry produces the over-the-shoulder lift, which is the more absurd of the two and the one people tend to send first. Each template ships with separate wording for one-photo and two-photo input, and the prompt stays editable, so you can set the location, the energy and - the edit that matters most - who is carrying whom.
You do not need a photo of the two people together. The tab opens in two-photo mode: upload a portrait each and the model places both into one scene with matching light and perspective, which is the whole point if the two of you live in different cities. Single-photo mode works differently from what people expect - its built-in prompt reads 'someone carrying the person in the photo', so the AI invents the person doing the carrying. If your single photo already contains both people, rewrite that line to describe them and the model will use what is in the frame instead.
New accounts start with 10 credits. In two-photo mode the picker runs on the native effect roster of 17 variants across 5 families - Runway, Kling, PixVerse, Vidu and LTX-2 - where Runway Gen-3 at 720p for 5 seconds costs exactly 10, so the starter balance covers one complete clip. Single-photo mode opens the full image-to-video roster of 44 variants across 15 families, with Seedance 1.5 Pro at 8 credits for the same render. Free-tier clips are watermarked and download is disabled; a paid plan removes both, which is the version you actually want before dropping it into a group chat.
How it works
- 1
Choose one photo or two
Two-photo mode is the starting state and takes a separate portrait of each person, compositing them into a shared scene. Single-photo mode takes one image and, by default, asks the model to add the second person - edit the prompt if your photo already shows both. Single-photo mode also unlocks the wider model picker.
- 2
Upload your photos
Clear, well-lit shots where both faces are visible and not too small in the frame. Full-body or three-quarter photos give the model a better sense of height and posture for the carry, which is what makes a piggyback read as a piggyback rather than two people standing very close together.
- 3
Pick the carry
Piggyback Ride for the joyful on-the-back version, Fireman Carry for the over-the-shoulder lift. Each loads a starting prompt matched to how many photos you uploaded, so the wording changes automatically when you flip the toggle.
- 4
Say who carries whom
Edit the prompt to name the roles and the setting - a beach, a hallway, a football pitch, a kitchen. Specifying who does the lifting is the single edit that most improves a two-person clip, and on a joke clip getting it backwards ruins the joke rather than just the composition.
- 5
Set model, length and ratio
In two-photo mode Runway Gen-3 at 720p and 5 seconds is the default and the cheapest at 10 credits; PixVerse v3.5 and Vidu Q2 Turbo are next at 18. In single-photo mode Seedance 1.5 Pro does the same render for 8. Switch the ratio to 9:16 if it is going into a story or a Reel - the tab opens on 16:9.
- 6
Generate, then send it
A few minutes later, check the arms and both faces. Free-tier clips are watermarked and stay in your library, which is fine for confirming the gag works; a paid plan removes the watermark and unlocks the MP4 so you can actually send it to the person who just got carried.
Why use this tool
Two carries with different comic timing
Piggyback Ride is warm and joyful; Fireman Carry is the absurd one. They are separate prompt recipes rather than two speeds of the same motion, so running both from the same photos gives you a real choice about which version lands.
No photo of you together required
Two-photo mode takes a portrait of each person and builds one scene around them, matching lighting and perspective. That is the case this tab exists for - two friends, two cities, one carry that never happened.
You direct who lifts whom
The editable prompt sets the roles outright, so the joke lands the way you intended instead of the model deciding. On a comedic effect this is not a refinement - which one of you ends up over the shoulder is the entire punchline.
No per-image charge, but the roster shifts
Credit price is set by model, resolution and duration, never by the number of uploads. Two-photo mode does narrow the picker to the 17 native effect variants, where the cheapest is Runway Gen-3 at 10 credits for 720p and 5 seconds; single-photo mode opens all 44 and its cheaper entries.
Draft cheap, then re-render
Run one pass on the cheapest model in whichever mode you are in just to confirm the carry reads correctly, then re-run the same prompt on a heavier family if you want the clip to look better than a joke has any right to.
A free clip you can preview
The 10 starter credits cover one full Runway Gen-3 render in two-photo mode. It plays back watermarked in your library with download disabled - enough to see whether the gag works, with a paid plan clearing both so it can leave the tab.
Frequently asked questions
What carry styles can the AI generate?
Two built-in templates: a joyful piggyback ride and a fireman-style over-the-shoulder carry. Each has separate prompt wording for single-photo and two-photo input, which the tab swaps automatically when you flip the toggle. Because the prompt is editable, you can also push it elsewhere - a different setting, a different energy, or swapping which person does the carrying.
Do both people need to be in the same photo?
No, and two-photo mode is what the tab opens in. Upload two separate portraits and the AI combines them into one scene, matching lighting and perspective so the pair looks photographed together. Single-photo mode does something different by default: its prompt asks the model to invent the person doing the carrying, so if your photo already has both people in it, rewrite the prompt to describe them.
Is the carry me generator free?
One clip is. New accounts get 10 credits, and in two-photo mode the cheapest available model is Runway Gen-3 at 10 credits for 720p and 5 seconds - the starter balance to the credit. Flip to single-photo mode and Seedance 1.5 Pro does the same render for 8, or ByteDance V1 Pro Fast I2V for 7. Free-tier output is watermarked with downloads disabled.
Will the video have a watermark?
Free-tier clips carry a VdoBloom watermark over the video, and download, right-click saving and picture-in-picture are all switched off - so a free render can be watched but not forwarded. A paid plan removes the watermark and unlocks the MP4. For a clip whose entire purpose is being posted into a group chat, that is the step that makes it usable.
Do I need an account to use it?
Yes, a free one, and it takes less time than choosing which friend to carry. It is how the 10 starter credits get attached to you and how finished clips stay in your library rather than vanishing when you close the tab. No payment details are required for the free credits.
How long does it take?
Usually a few minutes. Two people interacting is a heavier scene than a single-subject animation, so it runs slower than a portrait effect on the same model, and longer durations and premium families add more time. Generation runs server-side, so you can close the tab and grab the file later.
What photos give the best result?
Sharp, well-lit shots where both faces are clearly visible and reasonably large in the frame. Full-body or three-quarter photos work better than tight head crops, because the model can read posture and relative height - and on a piggyback, relative height is what sells it. Sunglasses, deep shadow across a face and very low resolution are the usual causes of a disappointing likeness.
Can I upload a photo of my friend?
Only if they are an adult and they are fine with it. A friend's picture is theirs to approve, even for a joke, and a joke is a weaker justification than most. Your uploads and generated videos stay tied to your account and are not published by VdoBloom. Generations are moderated and requests that trip those checks can be rejected.
How is carry me different from lift and carry or bridal carry?
Carry Me is the comedic one: piggybacks and fireman lifts, built for friends and siblings. Lift and Carry leans romantic and dramatic, with playful, dramatic and sweet moods and no fixed carry style. Bridal Carry is tuned for the classic wedding pose and includes a golden-hour sunset variant. Same underlying mechanic, three different tones.
The clip came out strange - what should I change?
Two-person scenes fail on arms and faces, and each has its own fix. Arms that pass through a torso usually mean the model could not tell who was in front - name the roles and the carry style explicitly in the prompt. Faces that drift mean the source was too small or too soft, so re-upload larger, sharper photos, or try single-photo mode with a picture of both people. If it is still off, re-run on a different family; they vary a lot in how they handle two bodies in contact.