Audio-first community on your own stack?

I spent majority of 2020 stuck in my basement (like many others), and I built a couple of communities - learnt a few things and built a simple product called Space.

The original funnel to bring people into the community was
Twitter -> substack/slack.

There was a bi-monthly newsletter (substack) and the same content was posted inside slack; every now and then there would be an interesting discussion and if 6+ people wanted to chat about it, we'd run a Zoom every now and then; but we'd only really get those 6 or so people to join, and most of the time video was off.

I had also been working on a drop-in audio app (yes, like clubhouse), and in July we swapped out zoom for this app (Space).

While Twitter still brought in new users to the community, we also had cross-discovery from others who were using Space for their other community, as a user on Space you could see other sessions that are scheduled/live in a discover feed and join them; this was interesting as the original idea was to have each community be its own silo with no cross-over (think shopify vs etsy). Our live sessions went from just 6 people to 30+, the community grew from 40 to 300...

With Space, the community owners own the data (export email, drip campaigns, etc). We've spent more time on opening things up and adding integrations, for example, you can hook Space to Slack to kick these off and then members from slack can just click and join. If the session is set to record, it posts the recording back into slack for others to consume async.

I also saw Discord change some of their positioning to focus more on audio; discord audio channels have always been this "oh yea we do that too" to more of "audio" as a first-class experience.

I'm wondering if I should keep investing in integrations and open it up beyond slack? Or if we should add text to Space, and let people create micro-communities? Kinda like a whatsapp group? So not a full-fledged Circle/Tribes but something simpler... maybe focused more on podcast type communities?

Should communities be audio-first or audio-focused; with text secondary; or will they always be text-first?