When the platform blocks the input your product needs: how do you scope the mobile version?

We’re building Speechara around spoken conversations and live transcription/assistance. On web and desktop we can support the audio paths people expect. On iOS, the privacy model does not give a web app the same access to system audio; in many cases we are limited to the microphone.

That creates an awkward product decision. We can ship a mobile version with a narrower promise, ask users to route audio through a workaround, or avoid pretending that it has feature parity with desktop. None feels ideal.

I’m curious how other founders handle a platform-specific limitation that changes the core experience. Do you make mobile a companion product, redesign the workflow around what the OS allows, or hold it back until it is honestly useful? Examples of clear expectation-setting would be especially helpful.