Portable restrooms: first-last of the empty pad, then the row of units
Two site stills. Agent first-last interpolates the drop. A lifestyle restroom film is a joke bid, not a quote.
Two site stills. Agent first-last interpolates the drop. A lifestyle restroom film is a joke bid, not a quote.
A site coordinator knows the pad: gravel, grass, the fence line, the generator already sitting there. The offer is appearance — empty ground, then the row of units you will actually deliver. That is a bounded motion problem. You already own both stills, or you can shoot them. Do not prompt a model to invent the festival.
Empty pad, then the row
Photograph the pad as the event will see it before drop — same lens height, same kerb or fence, same sky family. That is the first frame. Photograph the same geometry after the units are set, or from a previous drop on that site if the layout matches the quote. That is the last frame.
First-last frame takes those two stills and generates the motion between them. Ordinary image-to-video knows where it begins and improvises the rest. First-last has to land. The arc of the shot is decided by you.
The two frames have to be plausibly connectable in the time available. Same location, same lighting neighbourhood, same camera. A close-up of one door morphing into a wide of a different field is a warp, not a drop. Keep the pad.
If the last frame is a different property, the model will connect them. The connection is the artefact everyone notices — a cousin fence, a building that is not on the site plan, a unit colour you do not run. Event staff walk that pad. They will pause the clip.
Prompt the travel, not the objects. The stills already contain the empty ground and the row. Name the mechanism: hold, then units arrive into the marks; camera locked; no new building. If the endpoints share geometry, the morph reads as a drop.
The agent interpolates the drop
Create a first-and-last-frame transition video is the named agent job: two photos in, one transition out. Attach a start photo and an end photo. The agent passes them as first and last frame and picks a first-last-capable model. Cost is priced per model, shown before you confirm. This is not the plain "turn one photo into a video" capability.
The AI video generator is the launcher when you are picking the row yourself. Either door still needs two stills. One hero of a single unit and a camera-move prompt is ordinary image-to-video, not a drop.
Flux 3 First Last Frame to Video is a catalog row for that interpolation: start frame and end frame, native audio, 720p or 1080p, durations from 5 seconds through 20. The listed rate is 9 credits per second. Audio is on — write the mix (a truck rumble, a quiet pad) or write silence. A silent brief still gets a soundtrack you did not choose.
Size the interpolation to the change. An empty square to a three-unit row does not need 20 seconds. Prove the pair at 5s or 6s. Climb only with a reason that survives a pause on the unit colour and the door side. A loop with the same still in both slots is ambient motion on one plate. This offer needs two states.
A lifestyle restroom is not a quote
A sunlit luxury bathroom, a marble stall, a smiling couple "at the festival" is a joke bid. The person reading the packet is matching your clip to a gravel pad and a unit count. If the clip shows spa lighting, you are not in the packet. You are in a different industry.
Do not:
- Text-to-video "premium portable restrooms at a beautiful wedding." That invents the venue and the SKU.
- Generate a host explaining hygiene. That is a talking head you do not need, and it is not the drop.
- Swap in a last frame from a stock luxury trailer you do not own. The model will land on a unit you cannot deliver.
- Invent ADA geometry the quote did not include. A ramp that appears in the interpolate is a claim.
The shippable clip is the pad you photographed, the row you photographed, the interpolate between them, and a line you typed after lock if the count or the date window has to sit on the file. Overlay is authored type, not a generate. The units do not have to talk.
Equipment rental is a different object
AI video for equipment rental companies is teach-the-tool: start, operate, shut down a machine so a first-time renter books. That is a nearby audience. It is not a restroom drop. Copying a skid-steer demo onto a bank of units is how a site pad becomes a how-to you are not licensed to give.
Your object is the pad and the units. Do not generate the ballroom. One still of a unit on a white sweep is a product turn, not a site drop. A last frame with more units than the quote is a count you will be asked to honour.
Would you attach both stills to the bid? If no, stop. If yes, run the agent, watch the fence line, and ship the interpolate.
FAQ
Can I animate one photo of the empty pad and hope the units appear?
You can. The ending is then the model's guess — count, colour, door side, a building that is not on the plan. Pin the row as the last frame or you are not selling the drop.
Is the agent the same as uploading one image to the video generator?
No. One image is open-ended motion. Two images is a bounded landing. The agent notes that this capability needs first frame and last frame. Missing the row of units is a different job.
Why not a cinematic wedding-restroom commercial?
Because the packet is a site. A marble stall is a unit you do not run on that pad. First-last of the real geometry is the quote. Lifestyle is a joke bid.
Do the two photos have to be shot on the same day?
They have to share geometry. Same pad, same lens family, same lighting neighbourhood. A noon still into a night still is a time-lapse you should prompt as one. A different venue is a morph.