It seems like there's not much market in functionality or tools, more just adventures and adventure content (maps, beasties, whatever). I've never touched virtual gaming beyond being nudged into using Roll20 when a meet-up wasn't possible (very limited) and Foundry (better). But for me I like to game with friends. Otherwise it's spending tens of hours by myself writing something for a bunch of strangers to enjoy and then fuck off into the ether once more. But the tech interests me.
There isn't much market in Functionality or Tools because the cost/benefit is never going to be economically viable given the market size. These are all free/community/hobby projects for a reason.
The other issue is that, outside of an adventure pack where all the tools and how they need to be used are heavily bounded, any extension beyond the most basic will need so many customization options to fit potential use cases you're often not too far removed from the scripting language.
I use MapTool as my VTT and do a lot of my own scripting/coding. I had started down the path of coding up little character sheets that would pop-up and autocalculate HP/defenses with an eye to automating things like targeting and damage rolls. Mid way through I realized I was spending significantly more time trying to provide not just error checking for HP values but for making sure that mistakes could be easily amended or rolled back than I was just updating values on the Characters and just adding a "click these numbers to do damage"
If you were looking to make money on these sort of thing, you'd be much better served by going Whaling:
That is, make a some custom modules as a 'portfolio' and then approach the big Streamer DMs and offer to customize their VTT to their specifications.
You'll be competing with their cabal of simps willing to do it for free, but anyone deep enough into business of streaming will know sometimes Free labor is the most expensive.
Even then, I don't think you'd ever get to make that your day job. I think at best it'd be beer-money sort of hobby.
Additionally, a LOT of what they'll want will be art based (either direct graphics or at least prettifying the UI) so unless you are also a digital artist, you'd need to have a artist you can use and they'd take a cut.
The only other market I could see, and I hate to even suggest this because I hate everything about it, would be AI DM integrations. Plugins to allow an AI DM to interact with the VTT and assets. Because as much as it fills me with disgust for the world, there would almost certainly be a viable market.
The only thing that doesn't make me full mald is that LLM AI is just hallucinating so couldn't coherently interact with the hard reality of things on a (virtual) table top, so you'd need a traditional AI to handle enemies/board/orchestration, a strategic AI (I forget the term. I want to say bayesian, but might be neural net) to handle tactics if you don't want it getting stale, and a LLM for dialog.
to supporting tabletop TVs with overhead cameras that can track physical minis on the table,
Quick aside on this, having followed this stuff with interest since the early demos with with MS surface:
- the overhead cameras have tons of issues especially the more you try to farm out to them The promo videos don't show the AI vision going utterly insane on a fairly regular basis unless everyone is very careful and intentional with their movements.
- The most robust system I have seen for this was using a little bit of everything: RFIDs for the minis to track positioning, overhead camera tracking to determine what is being interacted with, and touch integration for the menus.
But even then, there was very noticable jitter. It looked flawless with the creators working it, but then the minute a normie stepped up "Ok whoa hold on you're moving the mini too fast/lifting it too high/needs to be kept over the table" and other such issues quickly cropped up.
- using RFID is the most reliable way if you want interactive minis, but using RFID is not compatible with "just put a TV on is back" unless you are OK with the RFID reader circuits being visible or serious hardware mods with the giant LC display.
- The map size was also fairly limited, and in both expect and subtle ways. it wasn't tiny it was at least 18x30, but when physical minis are involved, even just scrolling the map over a couple squares is suddenly a big deal