Audio (and maybe visual) Toy

Looking for advice on what hardware/parts to use, any pitfalls, or general advice/similar projects to look at.

I'm looking to create a little noisemaker which replicates the capabilities of the Gaia Memory toys from the series Kamen Rider Double (see; however, I want to combine all of the toys into one, small package (definitely can't just throw a waveshield in there!)

In other words, my project goals are: -Has a speaker and is able to play wav files (they don't have to be loud or high quality) -has an actual (male) USB plug (possibly, able to be used to modify the program later or at least to supply power) -Able to navigate between "sound fonts" or modes via a pair of buttons ("forward" and "backward") -each sound font is a set of several sounds -have it play one sound when a 3rd button is pressed. -I would also like to add a small LCD or OLED (I was thinking Ada's 128x32 OLED) to display a bitmap image which corresponds to the sound font/mode.

Additional goals which I am trying to achieve, but I'm not sure they are (all) possible: -size: 82mmx28mm rectangle, 14mm thick (roughly 3.25"x 1.125" x 0.5") with a USB plug hanging out one end. -battery-powered (no, it doesn't have to last for days or even would just be nice to not ALWAYS have it wired to a power supply) -USB charging

If needed, I was thinking that the OLED could be used only when the project is plugged into a USB port. Then, the battery would only be used to play audio (which should ease off the power requirements quite a bit, I think).

Thanks for your time, any comments are greatly appreciated.