For apparel and textile labels
Phone photos of the piece in. A month of Instagram, Facebook and LinkedIn posts out, captions written.
Upload up to nine photos of the real garment. Latent preserves its shape, colour, weave and the text on your label, then generates a month of posts with captions for Instagram, Facebook and LinkedIn. Free while in beta, no card.
Inpaint
One wrong shadow used to mean staging the whole set again. Paint over the part that is wrong, say what belongs there instead, and the rest of the shot stands. The fold you spent the morning on is still the fold in the picture.
What an inpaint leaves aloneReference photos
Prompt-only generation invents garments and drifts the colour. Latent starts from yours instead. Attach up to nine photos of the real piece, shot on your phone, and every picture is steered by them: same weave, same colourway, same stitch at the cuff. Only the scene around the garment changes.
How generation reads your photosA drop at a time
A drop is a full set of pictures, and the feed still wants something on the days between drops. Start from one prompt and the assistant drafts the month with you on the days you do not know what to post. The looks change the world around the garment, folded on oak to hung on a rail to flat on linen, while the reference photos hold the weave still.
What the assistant drafts The looks a garment can sit inEvery shape
A feed post, a grid tile and a story are three shapes of one photograph. Latent extends the frame outward instead of cropping into it, so the hem is still in the picture in all three.
Captions and publishing
The caption arrives with the picture, written to match it and written again for each account. Instagram, Facebook and LinkedIn each get their own words and their own shape. You read every caption before it goes out, and nothing publishes without you.
Where the posts goInstead of the patchwork
Single-picture tools stop at the picture. The set still has to be captioned, resized and sent, in whatever tab does each job.
Today
With Latent
One prompt makes the picture, the caption and the post. One browser tab, nothing to install.
Still images and captions, out to Instagram, Facebook and LinkedIn. Product and flat-lay stills, not people wearing the clothes. No video and no Reels yet. More places to publish are the road from here.
We used to reshoot the whole set over one wrong shadow. Now I paint over the shadow, say what belongs there instead, and everything I already liked survives it.
Free beta
Make the pictures, fix the one wrong thing, and send the posts. Everything you make stays yours.
Questions
Yes. Choose one look and generate the whole drop in it. Latent takes up to nine reference photos per product, and every picture is steered by the same set and lit the same way, so the grid reads as one set instead of a run of shots taken on different days.
No. Latent makes product stills: the flat lay, the folded stack, the piece laid out on a surface. It does not generate people, so a shot of someone wearing the piece is still yours to take. Everything else the drop needs, Latent makes.
Rules differ by platform and they change, so check each platform's current policy before you post. On Latent's side the facts are these: images are generated from photos of your real product, you review every image and caption before publishing, and nothing goes out without you.
Only you. Reference photos go into your private library. There is no public gallery and no shared feed.
Yes. Latent is in beta and free to use. No plans, no prices, no card. And it asks before it spends anything.
No. Latent runs in the browser. Photograph the piece, type a prompt, review, publish.