Launched this week

PackshotForge
Studio product photos, generated inside Photoshop
6 followers
Studio product photos, generated inside Photoshop
6 followers
I built this because I was tired of choosing between expensive product photo shoots and monthly AI photo apps that charge per image. PackshotForge is a Photoshop script: open a raw phone photo, pick a scene, and Google's Gemini model rebuilds the background while keeping your product (shape, colors, label text) intact. It lands as a layer. It runs on your own free Gemini key, so no subscription. 21 presets, marketplace export sizes, batch mode. Happy to answer anything!




Correction on pricing, since I got this wrong above and in the launch thread: Gemini image generation is NOT covered by a free tier. It's billed per image on your own key — roughly $0.02–$0.08 depending on the model, about half that in batch mode. I checked my own billing to confirm. What stands: no subscription and no markup from me — the cost goes straight to Google. Sorry for the sloppy wording; the landing page and listing are corrected.
Loving the no subscription angle and using my own Gemini key is smart. One thing that would help a lot: add a quick mask preview so I can confirm the product edges and label text are preserved before committing the background swap. Sometimes labels get slightly distorted and I'd rather catch that upfront than after the layer renders.
@nazmiyebaysytm Great suggestion — a pre-commit mask preview is now on the v1.1 list. In the meantime, the generated background lands as its own layer, so you can toggle its visibility to diff against the original and catch any label drift before flattening. Thanks for trying it!
Ran a quick test on a messy desk shot of my coffee bag and the label text stayed sharp while the background got cleanly rebuilt. Layer drops right into Photoshop exactly as described.
@sabrik94448 Thanks for putting it through a real messy-desk test — that's exactly the target use case. If you hit a scene the 21 presets don't cover, the custom prompt field takes free-form scene descriptions too.
the fact that it drops the generated background as a layer in photoshop is honestly so smart, you keep full editing control instead of being locked into whatever the ai spits out
@srandrrdxp That was the core design bet — Photoshop users don't want a black box. Everything lands as a normal layer with a mask, so you keep full editing control. Thanks!
Ran a quick test with a coffee bag photo and the label stayed sharp while the background swapped cleanly to a marble counter, way better than the AI photo apps I keep canceling after a month.
@mzeyyenxgtv That monthly cancel-resubscribe cycle is exactly why the script itself is now free and runs on your own Gemini key — you pay Google directly per image and nothing to me (correction: image generation is NOT free-tier — about $0.07/image, half that in batch). Thanks for testing it on a real product!
v1.1 is live — free update for everyone. The big one: Pixel-Lock (on by default). Your original product pixels are composited back over the generated scene through a Select Subject mask, so the model physically can't redraw your label or product surface. The result dialog also gained a Show original / Show result toggle — Nazmiye, your mask-preview suggestion pushed this to the top of the list, thank you! Also new in v1.1: multi-preset batch (every photo × every selected preset, with a batch report), transparent product cutout export with verified alpha, an async batch mode at half the per-image price, and an optional Stability provider. Cost stays honest and pay-as-you-go: you pay Google directly, about $0.07 per image full quality, half that in batch.
One more update: PackshotForge is now free. The script was $29 — it's now $0 on Gumroad, same link, no strings and nothing gated. If you bought it earlier, thank you sincerely; I'd rather have people actually using this than sitting on a paywall while I figure out where it goes next. Running costs are unchanged and still go straight to Google on your own key (about $0.07 per image at full quality, half that in batch) — I take nothing either way. v1.1 with Pixel-Lock is what you'll get: https://joktnova.gumroad.com/l/p...