The more I become accustomed to voice input for my notes, the more I wonder about the optimal integrated workflow with Logseq. Has anyone developed a speech-to-text process alongside Logseq that goes beyond simple copy-pasting?
Not exactly developed but I sometimes use the Recorder App on my Pixel and just share the transcript to Logseq. It’s not perfect but does the job for me.
1 Like
Yep. That’s what many of us do. In my case I use Tana Capture for the voice memos, and once I have the curated output, if suitable, paste it using Markdown in Logseq.
But the “copy-pasting” strategy seems a bit old-fashioned… tbh.
MacOS has native speech recognition, and MacWhisper for example works well offline across macOS apps too.
1 Like
But, what is the Logseq role in that situation: copy-pasting directly from the transcript? some text parsing in between?
It directly takes the output of the transcription, you are dictating into Logseq itself. MacWhisper lets you setup a whole range of different AI models to process the transcript in between, but I haven’t done that yet.
1 Like
You could use the logseq api to add notes, i played a bit with it, it is not perfect, but would be a way to use it. Main problem was to control, when to break a text into peaces.
1 Like
Anyone using open source android recorder for transcription that is then shared to logseq?
I just discovered that the android logseq app has a recording button in the bottom left, but it just saves the recording. I’m fine with that if there would be something like a background task that adds a transcript below the recording 
There is a quite nice open source program for voice input. Search cjpais/Handy on github (can’t insert links). It records a phrase, then converts it to text and pastes at cursor position. Nothing special to logseq, but at least no copypasting)
The new Logseq IOS app that works with Logseq DB has voice to text transcription feature.
3 Likes