In Review
13Under consideration
Influence image with image
I like the image to image when you can add multiple together. On some platforms you can enter 4 or more and adjust each one like edge to image, depth to image, pose to image, style ref, content ref, character, copy etc...adjusting the weights is fun, but I don't have my pirate diff models....definitely a UI box to do it , but I really hope it's in the cards
My Images bugs / improvements
if you select anything in the concept dropdown, it works, but afterwards any other selection will result in no images shown the task dropdown does not seem to work at all, nothing is shown under any selection except "task.." itself Select all checkbox actually works as an "invert selection", meaning if some images are already selected, they get unselected and vice-versa if you check the select all, then changes the dropdown so that no images are shown, select all auto-uncheck, but when you change the dropdown again, the images are still checked but select all remains unchecked. super weird bug if you scroll all the way down on any page THEN back up and change any of the dropdown values, all the images vanish and there is no way to make them come back without full refresh F5 Minor bugs / aesthetic URL says "archive", top menu says "My Images", page title says "My Renders" The "thumb size" slider should be percentage based on the image width, so that fully to the right shows the entire image, and it gets scaled down the further you go to the left (that way you don't end up with a lot of empty space if the image isn't wide enough)
Disable left and right arrow keys to change image when typing in text areas
After a render, when my cursor is in the "make this" box and I use either my left or right arrows key, the image changes and I lose all of the text in the box.
more archive sorting / navigation options: a way to sort by reverse date, or by month, so i can see the oldest first
Bug: organize button just leads back to the 15-minute message
Bug: in Archive, after selecting and deleting some images, then selecting more images, the Organize button just leads back to the 15-minute message, it doesn't reset to allow organization options. Fairly new bug?
add "seed travel" controls fine tuning to control the change rate
Allow a user to set a favicon / logo per space
inability to change it from “Fresh” and inability to change user photo / community banner and favicon is a bit of a inconvenience
More ways to manage renders, drafts
Can there be a way to select ONLY images in my personal gallery which are NOT organized into a project folder. Anything that’s essentially in our gallery’s root folder. Why: I’ll usually have a ton of bad renders from my workflow, and I’ll only keep select ones, which are organized into project folders. Anything outside of those are ones I want to delete, but if I select all, it also selects ones I’ve designated to keep. Being able to select unorganized images would allow for speedy bulk deletion of trash, rather than having to select one at a time or reply to each message on telegram with the delete prompt. Barring that, maybe an icon to indicate something IS in a folder somewhere on the thumbnail? Manual deselection from a select all isn’t that bad. Or even maybe a button on the main page next to select all that allows deselection of anything that’s been put into another folder. There’s probably a ton of ways to implement what I’m suggesting, but anything that allows image management without having to click each and every image without knowing if it’s been organized or not.
Select from render & upload in editor
Control net in the new unified editor only wants input from uploads. Polly gave me some contracode work around but I'm sure everyone would be happy if it was right there same as uploads like on other pages; recipes, in paint. 2 upload external files in editor. It's doable but wonky going through the recipe page for uploading everything. I would think maybe the gradient cloud, that could be a feature too, but I can see there might be reasons. More Buttons More access More easy broe thanks
Add API snippet on model pages
Like how we have copy codes for Triggers, but add a whole Json copy area or a different view for developers
In Progress
2Actively being built
Ability to batch process a bunch of images through a specific pre-determined process.
To illustrate, I want the equivalent of this: Basically: Upload/Select images. Set up the Stable2Go img2img/outpaint thing I want it to do. Press "Start" and it will go through each image automatically one by one, using the img2img/outpaint settings I chose before. This would make life much easier if I can just bulk-apply a repetitive task over multiple images, fully automated.
Internal server error when creating public space
To replicate: Set your username to Rock Lobster Write the bio "Just a lobster" Click Profile > check Create a gallery of public works > click URL Internal server error
Completed
6Recently shipped
Remember my prompt every time
The prompt field should always remember the last prompt by default, unless I wipe it
Tab gets temporarily refreshed with different tab content
When creating a My.Graydient account, not appear as a user in Stable2go
When paying by credit card with Chargebee (new way) with a new email, the Clerk login does not yet recognize the user
The ability to view all images across all your groups/spaces from a single cloud drive
For those of you have multiple groups to manage images, you might be having trouble finding those images when you come to Web UI This is why we've started building my.graydient.ai to give you easier access to your dashboard and other apps At the moment it only displays images created by Stable2go or your "main" login, and we'll be expanding this shortly
Bug: Archive showing duplicate photos
When scrolling down, the archive shows duplicated photos
Bug (model selection)
I’m using the latest version of firefox (and windows 10) (but I’ve also duplicated it in Edge) browser. The problem is that some times the browser freezes when in the fave model/lora selection menus. The blue bar (process completion) runs, but never completes. At that point its not possible to get a response from clicking. I’ve had to refresh the page to get out. It appears to be random,or some variable I’m not aware of. I’ve also started seeing a bug with the fav system. Attempting to fav or unfav a model drops you back to the creation page,without saving the fav or unfav. Thanks for your time.