Feature Request

Sometimes we have offline meetings that we record using voice recorder on ios. I'd like to be able to upload these audio files and have them transcribed just like a meeting. Alternatively if the iOS app comes out, give it the ability to record offline without an internet connect, then to sync to the server later.

Allow users to edit the title of each meeting, add a tag (#standup, #sales-call, etc), and perhaps allow a custom description would be great!

Would be awesome if users with privacy concern could opt-in to download a local LLM (Llama, R1,...), so nothing needs to leave the device!

Point to locally hosted LLMs

When playing back a call - can we have shortcuts (like on YT) to increase decrease speed at .25x increments? At work, I want to refresh key points previously discussed before I hop into a followup call. But i dont have the time to listen at the true 1x speed. need to speed through but want to hear the audio so i dont miss any nuances / tone of voice

It'd be great to be able to rescue deleted audios. Sometimes one makes a mistake and delete audios that shouldn't be deleted. There's apparently no way to fix that mistake in Shadow.

Either make the menu bar icon removable or (better) remove the app from ⌘-⇥ when the window is closed.

A way to blacklist applications so that they are not recorded when a conversation starts. I use Discord in my free time, and every time I join a voice channel, the recording starts. Since I don’t need the function in this case, whitelisting/blacklisting would be perfect.

Current onboarding friction.. 1) it's not clear that speaker names are editable. 2) Can we get a hint beforehand that speakers are mergable? Take this scenario- speakers are: [🔴Unknown] [🔵Speaker 1] [Me] Unknown and speaker 1 were the same person. I wasn't sure though what would happen if i set both names as the same.. if i'd end up with: [🔴Matthew] [🔵Matthew] [Me] but thankfully and very cool, it actually did merge them. -- However, not knowing what will happen with actions I can take adds hesitation/uncertainty –– even when the feature is already built in and actually available.. without a cue/clue, a new user wouldn't know. so user onboarding friction [🔵Matthew] [Me]

I just came off an hour call. It picked the wrong microphone and…at the end I got "failed" Likely it didn't hear the record - but if I'm in a meeting and we're at minute 6? Internally you should be notifying me if nothing is heard/recognized. Just start a process that quickly checks after a couple of minutes a snippet of sound (or use something real time) and if NOTHING is being recognized, Alert the user.