Download
The notarized Mac app is free to download.
Free to download. Sign in inside the app to start your 14-day trial.
Every published build will be signed, notarized by Apple, and listed here with its checksum so you can verify what you downloaded:
SHA-256 (Murmurtale-1.1.5-b1026.dmg)
9c981974a23d772bc215b1e95deb7972dbc1bae9f57e82c2186b172ccf5a3852
Once installed, Murmurtale checks for signed updates about once an hour and updates itself.
If macOS ever says the app can’t be opened: open Finder → Applications, right‑click Murmurtale, choose Open, then Open again. You only do this once.
What it needs
- macOS 14 or later on an Apple Silicon Mac (M1 or newer). Intel Macs are not supported.
- English only for now, both cloud and on-device. Dialogue formatting follows US English manuscript conventions.
- About 700 MB of free space — the app plus the on-device speech model it downloads on first launch.
- The Mac app is free to download. On-device dictation is included free (5,000 words a month on a free account); the Pro cloud engine, the default and most accurate engine, comes with a free 14-day trial for new accounts. Sign in inside the app to start the trial.
Install in three steps
- Open the .dmg and drag Murmurtale to Applications. Launch it from the Applications folder — not from inside the disk image — so macOS treats it as properly installed.
- Grant the two permissions it asks for on first launch: Microphone (to hear you dictate) and Accessibility (to place finished text at your cursor). That is the whole list — if a dialog asks for anything else, it is not Murmurtale.
- Let the speech model download once. Murmurtale fetches its on-device model (about 632 MB) and compiles it for your Mac’s Neural Engine. A minute later the quill appears in your menu bar and you are ready to dictate.
The Help Center walks the whole path in detail: Install Murmurtale on your Mac, then Grant Microphone and Accessibility and Your first dictation.
While you wait
See what the download gets you: the before-and-after on the home page shows raw speech becoming a formatted scene.