Oxygen-concentrator OEM: chassis still, not a cannula on a generated face

Q: Should I text-to-video "our concentrator rotating on a bench" and skip the photo?

You can. You will get a box with casters. It will not be *the* box often enough to run as the listing. The product video generator exists because the still is the spec. Image-to-video from that still is the motion job.

CPAP OEM already holds mask-and-hose. This listing is the concentrator chassis. Do not generate a wearer or a treatment VO.

CPAP OEM: mask-and-hose SKU hold, not a clinic sleep study already owns the interface: cushion, headgear, tubing, no generated sleeper. A concentrator is a different catalog row — cabinet, compressor, sieve beds, casters. A cannula on a constructed face is a patient you did not consent. Photograph the chassis. Leave the wearer off the generate.

The leftover SKU is the box

An OEM page answers which concentrator, which litre-class, which port, which caster set. The still has to hold the fascia you ship: the badge, the inlet filter, the humidifier port, the wheelbase. Text-to-video will invent a prettier cousin with a melted logo and a flow number the sampler liked. That is a failed SKU, not a cinematic upgrade.

AI product video is the door for that hold. Two or three clean angles of the unit you actually ship. Prompt the camera, not a bedroom. Do not prompt "a peaceful patient on oxygen in 4K." The CPAP post already named why that sentence is a person. This post is the same rule on a different object: the chassis, not the mask.

On-screen catalog strings are type you wrote. Add a text overlay burns the SKU, the litre-class, or the "continuous flow" line you typed — whole clip, top / center / bottom. A video model will not typeset a medical catalog number. Overlay after picture lock. If the still already letters the imprint, do not ask the generate to "clarify" it.

The pack shot is the contract. Motion is allowed. A titration night, a hospital bay, or a generated nasal cannula is not.

Photograph the chassis you ship

Use the current pack shot or a bench photo of this revision. Do not generate a cabinet colour, a caster geometry, or an imprint the SKU does not have. Wrong geometry on a medical device is a safety-adjacent miss, not a styling choice.

If you need motion, keep frame one as the file you sent. Image-to-video is the still-to-motion job. Seedance 2.5 is the family that row runs: four to thirty seconds, 720p, native audio with the picture, 47 credits listed. I2V is that still-to-motion job; the published URL is /models/seedance-2-5. Do not follow a folded Seedance I2V slug. Frame one is the chassis photograph.

Prompt only travel: a slow orbit that keeps the badge readable, a gentle pull along the handle, empty bench tone. No face. No bed that invites the model to grow a sleeper. A last frame is optional, and only when it is a second approved still of the same unit. Two competing cabinets is a morph. A last frame of a cannula on a face is a body.

If the listing is a chassis SKU, shoot the chassis. A "system lifestyle" paragraph will glue the wrong elbow onto the wrong cabinet. The CPAP post already holds the hose. Do not restage it here so the concentrator has a scene.

A cannula on a generated face is a patient

The cannula is not atmosphere. It is an interface on a person. A model asked to "show how the unit is used" will invent a wearer. That is a likeness you do not have a grant for, and in this category that person reads as a patient.

Do not first-last a chassis-only plate to a lifestyle still of tubing on a face. The middle of that interpolate is a body. Do not crop a generated sleeper and keep the tubing. Cropping a fake does not make it a product shot.

Minors are a hard stop. Do not generate a child on a concentrator. Do not "age up" a school photo. A real named person, filmed with a replica grant, is a human. A generate of that person with a cannula is not.

Treatment VO is a YPP mouth

YouTube's third inauthentic-content bucket is AI-generated personas that present as human experts on health, legal issues, finances, or politics. A constructed host walking litre-class or SpO2 talk is the same shape as a constructed doctor diagnosing. Labelling the face AI does not leave the bucket.

Native audio on Seedance 2.5 will invent a mix if you do not write one. Write silence, or write a quiet compressor rumble and no speech. Quoted treatment lines in the prompt do not become a safe soundtrack. They become a mouth you did not hire, delivering health talk you cannot stand behind.

A silent orbit of the cabinet is not that shape. The kill is persona plus prescription. Do not glue a talking protocol onto the product file. If a channel needs a voice, that voice is a real named clinician on a different plate, or it is type you typed. Health claims stay on the approved labelling. The clip can show the object. It cannot promise that a generated wearer's saturation improved.

FAQ

Is this the same as the CPAP OEM post?

No. That post is the mask-and-hose SKU hold: cushion, headgear, tubing, no generated sleeper. This listing is the concentrator chassis. Do not restage a cannula on a face so the cabinet has a scene. Photograph the box you ship.

Can I show a generated patient with a nasal cannula if I blur the face?

No. The cannula is still on a person, and in this category that person reads as a patient. A blur does not create a grant. Show the chassis. Overlay the SKU. Leave the wearer off the generate.

Should I text-to-video "our concentrator rotating on a bench" and skip the photo?

You can. You will get a box with casters. It will not be the box often enough to run as the listing. The product video generator exists because the still is the spec. Image-to-video from that still is the motion job.

If I prompt "no medical advice," can Seedance still put a wearer on the chassis?

The face is still a constructed presenter, and treatment talk is still health. Hold the cabinet. Write silence or a quiet rumble. If someone must speak, put a real named person on camera, or do not prescribe.