AskClaude/Voice
19 Jul 2026 - 19 Jul 2026
- Not even sure if this properly belongs in AskClaude, has nothing to do with Claude.
- Spec for voice plugin
- Direct voice->text (and yes this could use Claude)
- Make it easy to bring in Google Recorder recordings (from phone). This has its own ->text transformation.
- OK I want to turn this page into a Claude Code spec. Seems like that should be trivial. Hm it is so stupid to have separte text and code worlds!
- Command: pass to claude code as a spec, optionally specify an exiting project (repo).
- Google Recorder sucks, it insists on turning "fucking" into "f" like this is some corporate memo instead of my own thoughts. Need a subtitute, I bet Claude itself would work, what are the alternatives (for getting voice into Logseq)
- Actually Claude itself works, if you don't mind it trying to answer things
- Ask Claude
Voice-to-Logseq Options
- Claude directly - Doesn't have a mobile mic input, so not directly useful for quick capture.
- Better transcription apps that won't bowdlerize you:
- Whisper (OpenAI) - runs locally or via API, no content filtering, excellent accuracy. Apps like Whisper Transcription (iOS) or Aiko use it
- AudioPen - transcribes and lightly reformats, no censorship issues
- Otter.ai - doesn't filter profanity
- Notta - same
- Local/self-hosted:
- whisper.cpp on Android via Termux if you want full control
- Faster-Whisper desktop setup
- Direct Logseq integration:
- Logseq has a built-in audio recording feature but you still need transcription
- Logseq Whisper plugin exists - records and transcribes in one step, no filtering
- Probably the cleanest solution: The Logseq Whisper plugin if you're on desktop, or Aiko/Whisper Transcription for mobile capture that you paste in.
- Google Recorder's profanity filter is opt-out on some Android versions - worth checking Settings → Recorder → if that option exists on your device.