Is there any FOSS full speech recognition user friendly software yet?

Today somebody in a group I'm in which has some accessibility issues was yet again complaining that their Dragon Speaking software was not playing nice with Firefox, which led me to see if there was an alternative, and surprisingly i found none workable at the plain user level beyond Dragon, and upgrading for that person might actually be costly (From what they say it starts at nearly $200 but apparently can go as high as $700? Not clear yet).

So, obviously now I'm checking about the FOSS side of things, a search has been inconclusive as i see stuff for developers, multiple different projects (which is a marked improvement from a decade ago when i last tried and failed to do this), but so far haven't found anything at the user level.

Have i overlooked something? Or is it that we're many years later still at the "building libraries" stage without actual user-level stuff people can just apt-get or download?

Quick edit: I must insist, is there something for USERS, not DEVELOPERS, that i have overlooked? APIs or commandline programs or learning models are not a software i can hand to my non-programmer friend to install on their computer to replace Dragon to help them write on Firefox