AI character card browser and editor
Budget: $30 – $250 USD
Many text-based AI programs utilize character cards for loading chat personalities and RPG scenarios. Though there is an informal standard for these character cards, no offline software exists to manage or create them. I would like a simple platform-independent character card browser and editor. Python is probably sufficient, but I am willing to consider the use of any other programming language that is freely usable on windows/linux/mac.
Character cards consist of a PNG image and JSON character information. The JSON is embedded in the PNG metadata (somehow). The full character card V2 spec can be found here: https://github.com/malfoyslastname/character-card-spec-v2
An example of a web-based character card V2 creator can be found here. This should give you an idea of how exactly the character metadata is extracted/merged with the PNG, as well as provide an example of all the optional fields in a V2 character card: https://github.com/gaffe-buck/tavern-v2-character-creator
I have attached two examples of character cards - both the raw JSON character info, and the PNG with the JSON info merged in.
Behavior of the application should be as follows:
The UI should consist of a resizable window primarily consisting of two panes. At the very top should be a browse button to select the directory to scan and display character card PNGs from. The left pane (roughly two thirds of the window, adjustable) should display a gallery of character cards found in the folder indicated. These should be ~80x80 thumbnails and should be labeled with the name of the character, extracted from the character info in the PNG. The right pane should display all the character information fields for the selected character in editable text boxes. At the bottom of the right pane there should be two save buttons: "Save overwrite" and "Save new". Save overwrite should save the edited character details into the selected PNG. Save new should open a dialog box to select a file name and location to save a new copy of the character card with the edits. A third "Create new" button should clear out all the character info fields. A 4th button labeled "Add/change image" should allow the user to add an image to a blank card or replace the image for an existing card. A quick mockup of what I'm looking for has been attached.
Character cards consist of a PNG image and JSON character information. The JSON is embedded in the PNG metadata (somehow). The full character card V2 spec can be found here: https://github.com/malfoyslastname/character-card-spec-v2
An example of a web-based character card V2 creator can be found here. This should give you an idea of how exactly the character metadata is extracted/merged with the PNG, as well as provide an example of all the optional fields in a V2 character card: https://github.com/gaffe-buck/tavern-v2-character-creator
I have attached two examples of character cards - both the raw JSON character info, and the PNG with the JSON info merged in.
Behavior of the application should be as follows:
The UI should consist of a resizable window primarily consisting of two panes. At the very top should be a browse button to select the directory to scan and display character card PNGs from. The left pane (roughly two thirds of the window, adjustable) should display a gallery of character cards found in the folder indicated. These should be ~80x80 thumbnails and should be labeled with the name of the character, extracted from the character info in the PNG. The right pane should display all the character information fields for the selected character in editable text boxes. At the bottom of the right pane there should be two save buttons: "Save overwrite" and "Save new". Save overwrite should save the edited character details into the selected PNG. Save new should open a dialog box to select a file name and location to save a new copy of the character card with the edits. A third "Create new" button should clear out all the character info fields. A 4th button labeled "Add/change image" should allow the user to add an image to a blank card or replace the image for an existing card. A quick mockup of what I'm looking for has been attached.