113 catalog models will not run without a still

/without/a-camera is a route set over input flags (photo-in, video-in, text-only, lipsync), not a synonym for "use Versely."

The 113 in this title is the photo-in class: models that set requires_image. They will not start without a still. They are not "Versely." They are one partition of the catalog. If you have nothing to upload — no camera, no phone footage worth using, no still on a drive — those 113 rows are closed. The without page exists because that closure sends you down a different set of routes than "I have a pack shot" or "I have a face I will not show." A homepage pitch is not a route.

Four flags, not a marketing synonym

Every active model in models.json declares what it requires. /without cuts on those flags. The source that named the hub documented the partition as:

• photo-in — requires_image — 113 in that snapshot: you must supply a still
• video-in — requires_video_input — 22: you must supply footage
• text-only — neither flag — 168: a written prompt is the only required input
• lipsync — content_type lipsync — 19: a face animated to speech (nine of them explicitly audio-to-lipsync)

113 + 22 + 168 = 303 because seven models require both a still and footage. The chips on the live page are counted at build from the catalog. The without source refuses a committed number that would go stale on a refresh. Trust the chips if they moved. Trust the flags either way.

Four input classes support six distinguishable route sets. That is why there are six /without URLs and not twelve. A page earns a URL when the missing input sends you down a different set of routes. "You don't need an editor" is the homepage. It is not a constraint page.

/without/a-camera is the empty-hands case: no shoot. The prompt is the footage. Pick a route by how long the finished video needs to be. It is not "use any of /models." Most of that index will not start.

The three routes that actually start from a prompt

The camera page only uses the text-only class.

Write the shot, not the shot list — one clip, today. Subject, action, camera, light, look. One continuous shot. Text-to-video cuts badly when a prompt asks for two locations. Set aspect before you generate. Take two or three of the same prompt before you change the prompt. Keep the take with the right motion. Hand-off: AI video generator, generate a video from text.

Start from a premise and let it break into scenes — 30–60 seconds with a beginning, a middle, and an end. One sentence first. Lock recurring characters, props, and locations as named assets before you generate. Scene-by-scene, then stitch, then voice and captions on the cut. A 6-scene piece with one broken scene is a broken piece.

Cover a voice you already have — you can talk, you cannot film. Cut the audio first. One generated visual per beat. Short clips, not one long one, so a bad beat costs one clip. Then caption.

Those are the routes. "Use Versely" is not one of them. If you needed the object you sell, you are not on this page.

Photo-in is a different missing-input page

The 113 (photo-in) are the models this title names. They are the opposite of a-camera. They refuse a blank start. Image-to-video, many lipsync rows, reference-locked product turns: still in, motion out. If you have a jpeg, you are not without a camera in the sense this hub means — you have a plate. Animate that plate. Do not rewrite a-camera to include them.

/without/a-product-sample is the commerce fork of the same flag: you need a product image before you need a product video. Generate or rebuild the pack shot, lock it as a reusable asset, then enter photo-in. A generated pack is an illustration. It cannot stand as evidence of a unit you have never seen.

/without/showing-your-face is lipsync and faceless routes — a different missing-input than "no camera." Without a script is the one constraint the catalog cannot absorb: not one text-only model needs no prompt.

Do not collapse those URLs. Photo-in, video-in, text-only, lipsync. Four flags. Six pages. A-camera is three text-only routes. The 113 are the door that stays shut until a still exists.

What a-camera does not fix

It will not show your actual product. A generated object resembles the category, not your unit.

It will not show a specific real place. Text-only models render a plausible "Brooklyn corner store," never yours.

It will not put you on screen. Generated people are not you. A founder video that pretends otherwise is a trust problem.

Those limits are why a-product-sample, a-studio, and showing-your-face exist as other URLs. The 113 photo-in rows are how you leave the text-only lane once you have a still. They are not a workaround for having nothing. They are the reason having nothing closes most of the catalog.

A single generation is a few seconds. Anything longer is several generations stitched together. That length constraint is the other honest line on the camera page. It is still not a synonym for the whole product.

FAQ

Are the 113 models I cannot use without a photo the same as "Versely has no text-to-video"?

No. Photo-in is requires_image. Text-only is the class with neither image nor video required. A-camera routes only through text-only. The 113 never start from a blank page.

Why doesn't /without/a-camera list every model?

Because most models are the wrong class. The page is a route set, not a catalog dump. Open /models when you already know the input you have.

If I snap one still with a phone, am I still "without a camera"?

No. You have a plate. That is photo-in: animate the still, or go to a-product-sample if the still is the SKU. A-camera is the case with nothing to upload.

Do the 113 include lipsync?

Some lipsync rows also require a still; some take video; some are premade avatars; one is text-driven. Lipsync is its own flag. Do not treat "113 photo-in" as "all talking heads." Showing-your-face is the lipsync route set.