Features

Two modes. One hotkey logic.

Hold Space to dictate. Double-tap to command. Murmly handles the rest.

mic

Mode 01 — Dictation

Transform thought to text with zero friction.

01 / AUTO-INSERT

Inserts where your cursor is

No separate windows or copy-pasting. Murmly acts as a virtual keyboard driver, injecting text directly into Slack, Notion, VS Code, or any text field instantly.

02 / REAL-TIME CLEANUP

Cleans up as it types

Say "um" and "ah" as much as you need. Our real-time post-processor removes disfluencies and fixes grammar while preserving your unique speaking style.

03 / APP AWARENESS

Knows the app you're in

Context is everything. When you're in VS Code, Murmly understands PascalCase and camelCase. In Gmail, it defaults to professional correspondence formatting.

04 / BUFFER CONTROL

60-second undo

Made a mistake or changed your mind? A quick double-tap of the escape key rolls back the entire insertion and clears the buffer instantly.

05 / BIOMETRICS

Voice profiles

Advanced vocal fingerprinting ensures that only your voice triggers the dictation. Background chatter in a coffee shop? Completely ignored.

system_init // loading...
claude_sonnet // ready
> murmly // push_to_linear_and_notify_slack
terminal

Mode 02 — Commands

Execute complex workflows via intent.

01 / SEMANTIC INTENT

Speak intent

Don't say "Click file, click save". Say "Send the last three paragraphs to Sarah on Slack". Murmly understands the goal, not just the action.

02 / LLM BACKBONE

Routes through Claude Sonnet

Powered by state-of-the-art reasoning models to ensure high-accuracy mapping of verbal commands to complex API sequences.

03 / SAFETY FIRST

Asks before it acts

For high-stakes actions like bulk deletions or emails to clients, Murmly provides a visual confirmation toast before finalizing the execution.

04 / RECOVERY

60-second undo

Accidentally archived a thread? Command 'Undo that' within 60 seconds to revert the system state back to its original position.

05 / EXTENSIBILITY

MCP under the hood

Built on the Model Context Protocol (MCP), Murmly integrates with any local or cloud tool that supports standardized AI orchestration.

How it stacks up

FeatureMurmlyWispr FlowOS Dictation
System-wide Latency< 150ms~400ms~800ms
Contextual Commandscheck_circleLimitedcancel
Undo Window60 SecondsImmediate onlyNone
Privacy EngineOn-device firstCloud-basedOn-device
Ecosystem

Works where you do

descriptionNotion
codeVS Code
mailGmail
forumSlack
grid_viewLinear
terminaliTerm2
edit_noteObsidian
format_shapesFigma

Frequently Asked

How secure is my audio data? expand_more
Murmly uses on-device transcription for the primary dictation mode. Audio is only sent to our secure, SOC2-compliant cloud for complex command reasoning, where it is purged immediately after processing.
Does it support other languages? expand_more
Currently, Murmly is optimized for English, Spanish, French, German, and Mandarin. We are rolling out support for additional languages throughout the year.
What hardware do I need? expand_more
Any modern PC with a built-in or external microphone. For best results with the Voice Profile feature, a cardioid microphone is recommended but not required.

Start speaking today.

Join 10,000+ professionals using the voice layer for modern work.

BUILD 2.4.0 // LOCAL WHISPER V3