Disclaimer
Is this plugin open source? Yes — MIT licensed
Is this plugin completely free? Yes, but you need to supply an LLM provider API key
Is this plugin vibe-coded beyond the author’s ability to comprehend how it works? No
Community Directory: https://community.obsidian.md/plugins/voice-append
Hi everyone,
I often think of an addition to an existing note while walking. Opening the note on my phone is easy, but typing a useful paragraph is not. A normal voice-note app also creates another inbox that I have to process later.
I built Voice Append for that specific workflow: open an Obsidian note, record the next thought, and append a cleaned-up version directly to the same note.
What it does
- Starts recording from the note, mobile toolbar, ribbon, or command palette.
- Saves the recording locally before network processing begins.
- Retries interrupted or failed jobs, so you can record offline and process later.
- Uses configurable OpenAI-compatible providers for transcription and cleanup.
- Supports separate transcription and cleanup providers when needed.
- Lets you edit the cleanup prompt and optionally provide note context and familiar terminology.
- Can generate a filename for a new note whose body was empty before recording.
- Works on iPhone and desktop, with English and German interfaces.
There is no Voice Append server. Audio goes to the transcription provider you configure. The transcript—and optional note context—goes to the configured LLM provider for cleanup. Provider keys are stored locally using Obsidian Secret Storage.
The recordings view lets you retry failed jobs, download their audio, change their target note, or delete them.
Links
Community Directory: https://community.obsidian.md/plugins/voice-append
Source code and documentation: https://github.com/wko/obsidian-voice-notes
Bug reports and feature requests: https://github.com/wko/obsidian-voice-notes/issues/new/choose
The README contains a short demo of the complete recording and append workflow.
This is the first public release. I would especially value feedback from people who capture thoughts on mobile: where does this workflow still create friction for you?