Photoshop to Spine 2.0

SvissBro

Remember the old PhotoshopToSpine script?
It does its job perfectly, but... what if we took it to a whole new level?

Imagine a Photoshop to Spine 2.0 concept - a tool with a full-fledged visual interface, advanced asset size optimization, Auto-Bones, Auto-Meshing, texture auto-rotation for tighter atlas packing, a Changes Tree, and so much more.

What’s under the hood?
Besides the core features of the standard script, I've added:
🔍 Smart Deduplication: Automatic detection and replacement of duplicate textures to save memory.
📊 Atlas Stats & Previews: No more "blind" exports. You instantly get a visual preview of your future atlas and precise size statistics.
🌳 Changes Tree & History: Every single action and optimization is logged in a transparent history tree.
🎯 Canvas Origin: A convenient visual setup for your root point directly in the app.

And this is just the beginning!
I'm planning to add even more powerful features to automate the daily routine:

Skeleton Presets for a quick start.
Action Graph (a system to record and playback custom processing scenarios).
Batch Processing for mass export and optimization.
Mesh Tracer for perfect outline-based mesh generation.
2.5D Rigging (the concept I showcased in my previous post).
Smart Rigging - advanced algorithms for auto-generating bones, creating better meshes, and auto-calculating weights.
Smart Renamer for easy bulk renaming of JSONs, PSDs, and textures.
PNG Sequence Trimming (I’ll show this script a bit later 👀).
What do you think of this evolution of the standard Photoshop to Spine workflow? Would you like to see this kind of functionality in your daily pipeline? Share your thoughts below!
More info in my profile:
https://www.linkedin.com/posts/svissbro_spinescripts-scriptsforspine-activity-7442972669156683795-LK24?utm_source=share&utm_medium=member_desktop&rcm=ACoAADco5FsB_O_2esiM7WTpm8fJbV-tLi4EjeAhttps://esotericsoftware.com/forum/assets/files/2026-03-26/1774546604-205653-photoshop-to-spine.mp4

Nate

Thanks for sharing! From the video it seems like you have a LOT more functionality than just PSD -> Spine. Visual programming? It's hard to make much sense of what is going on in the video timelapse.

Smart Deduplication: Automatic detection and replacement of duplicate textures to save memory.

The standard way to do this is to use a Photoshop "smart object". You'll want to support those not just for deduplication, but it enables using Photoshop features (filters, etc) that can't be supported by 3rd party PSD parsing.

BTW, your post is obviously AI generated and that is perfectly fine, but note it may turn off some users. A lot of people are using AI to generate sloppy software that doesn't work well. Your software may very well be awesome, but if presented by AI generated text in the same way that others use whose software is not awesome, it can give the wrong impression.

SvissBro

Nate Thanks for the detailed feedback! Yes, the tool goes beyond simple PSD → Spine import — the idea is closer to a small visual workflow / action graph where rigging steps can be automated and reused. The timelapse probably wasn’t the best way to show the structure, I’ll share a clearer breakdown soon. Good point about Photoshop Smart Objects — supporting them makes sense, especially for compatibility with non-destructive workflows. I’m currently focusing on detecting identical textures to reduce atlas size, but Smart Object support is definitely something to explore. Regarding the AI text — English isn’t my native language, so I sometimes use AI to help phrase descriptions. 🙂

Nate

The main thing about smart objects is Photoshop includes a rasterized version. That way filters and all kinds of Photoshop nastiness can happen inside the smart object, but Spine or your tool can just use the rasterized version.

FWIW, Spine's packing detects duplicates. It has an alpha threshold setting so even slightly different images are treated as the same.

AI is great for phrasing and conversion to English, but if it includes overly hyped text that reads like AI wrote it, that's not good. Also if you let it format things with emojis and emdashes then it looks like AI. More and more people want to know a real human is behind a project, as there is a huge flood of projects that are AI generated junk and you don't want to be associated with that.