17 Feb 2026
Voice to Markdown: The Murmr Shortcuts Plan
How Murmr uses Shortcuts to turn on-device speech into structured Markdown for Obsidian and developer workflows without cloud integrations or tokens.
Murmr exists to make one loop frictionless:
Speak → structure → ship.
No accounts.
No tokens.
No background servers.
Just local speech, structured properly, and routed where it needs to go.
The key to making that work across iOS and macOS is Shortcuts.
The core idea
Instead of baking GitHub APIs, vault logic, or webhook management directly into Murmr, the app exposes a focused automation surface.
Murmr becomes:
- the capture engine
- the on-device transformation engine
- the structured Markdown generator
Shortcuts becomes:
- the router
- the integration layer
- the automation glue
That separation is intentional.
It keeps Murmr simple.
It avoids credential handling.
It reduces review and security overhead.
It keeps everything local-first.
The Obsidian workflow
Obsidian works on Markdown files.
That simplicity is the opportunity.
The winning loop looks like this:
- Press the Action Button.
- Record a thought.
- Murmr transcribes on-device.
- Murmr applies a template (for example “Obsidian Inbox”).
- Shortcuts appends the formatted Markdown to
Inbox.mdinside the vault. - Optionally open Obsidian.
No plugins.
No API keys.
No custom integrations.
Why this works
Murmr handles structure.
Shortcuts handles file operations.
Obsidian reads Markdown.
Each piece stays in its lane.
Example: Obsidian Inbox template
Raw speech:
“Need to refactor the Kanora AirPlay pipeline and test edge cases with multiple outputs.”
Murmr produces:
### 20:14
- Refactor Kanora AirPlay pipeline
- Test edge cases with multiple outputs
#voice #murmr