I don't work with data but are dataset formats really so similar that a single reading API can fit all of them, including future ones which may have strange and new forms of data we find useful?

Do you not also then have to worry about bugs in the web assembly tainting results produced from this data? its not like a managed library for reading where you can simply update the library if you find a bug. you would have to redownload the dataset which is bundled with the anyblox reader? or maybe patch the file which would be more involved. Im assuming these datasets would be large enough for this to be an inconvenience.