I don't mean a use similar to that of Wispr Flow or Typeless (which is what I personally use), but the possibility of loading the audio of personal audio notes, phone calls or meetings into Recall which I would then like to be able to transcribe and query with the LLM.

This thread popped up earlier today.

One use where I would use it.

When I search within a specific category in Recall and review the notes in both Reader and Notebook, I want to extract information or make notes from the data. This might work with Wispr Flow (et al), but having it integrated into Recall would provide a much smoother experience.

Is it essential, considering all the other speech-to-text tools that have popped up? Probably not, but having it in one integrated space would definitely be useful.

Having speech-to-text to control recall would also be useful. Adding tags by voice could certainly speak things up.

I'm sure there are many other uses that others will contribute.

Let me explain the use case. I use both Whisperflow and Superwhisper — great tools for converting speech to text, but that's not really the point here. The point is the transcribed text actually ends up.

For example, I journal every day, and throughout the day ideas or things I want to note just pop into my head. Right now that process is a pain. I have a text file sitting on a server, I have to open it up, and then I dictate into it whenever something comes to mind. Super annoying.

By the end of the day I've got a text file with 20 or 30 random phrases, ideas, and action items — and then I have to drag in some LLM to turn all those little snippets into something coherent: a proper summary, a list of action points, etc. It's extremely disruptive when your tooling is this fragmented.

It would be awesome to have a single place where you could just dictate individual notes for your second brain, and they'd be stored in an environment where you have all the tools you're already used to from Recall — right there, ready to go.

I am curious about the need for this feature, given tools like Wispr Flow. Please can you share if this is something truly needed, given the existing voice apps that are being widely adopted today?

agreed

Transcribing from voicenotes would be fantastic, to jot down ideas, thoughts and notes.

Yes, voice notes would be great

Sankari Nair voice notes with diary / project log would be perfect for professional use. As Project Manager I need to capture very often some key events, information, what happened during a day etc. Having easy way to capture notes with dates is like external memory feature :)

If voice transcription is to heavy for implementing think about Mac application - combined with other tool like superwhisper would be a first step.

I'd like to vote for voice notes too. Dictation can make capturing ideas, entering comments, questions and comments so much easier. Thank you.

I don't think it's fundamentally a bad idea. Personally, I don't believe I'd use this functionality because for me, Recall isn't a notetaking app in the sense that I just dictate whatever crosses my mind. It's a database for collected currated and structured information that I mostly gather from external sources. And when I do make my own notes, it's usually a research result from the AI chat. I'd very rarely want to dictate my own unstructured thoughts into it. If anything, I'd prepare them as a document and upload that document to Recall later.

That said, other people might have different use cases. But the challenge is getting the transcription clean and accurate. Ideally with formatting too, so there's not too much post-processing work. And that's actually worth its own application. That's the domain of dictation software that's really good at this and gets transcription factually 98% correct. Otherwise people won't use this feature in Recall. And that's where I see the problem.

Recall isn't dictation software. So they'd basically have to provide one under license, similar to how they provide transcription under license. But that's one of the most expensive token-consuming parts of the whole business case. That's why I have some doubts about whether this is a sensible requirement.

But in case you're interested in using proper dictation software, I'd at least have a suggestion here: https://wisprflow.ai/r?FERNAND3 it's a fantastically accurate voice transcription software and you can dictate in any field you want with nearly 98% accuracy with your own dictionary and many other features. It's running on the Mac OS, Windows, iPad, iPhone, and there is an Android version which is announced.

So using WisprFlow you can just dictate freely into your already existing Recall app on any supported device.