Local by default
Speech generation and supported voice workflows run on the Mac, with clear model setup and runtime state.
All workmacOS · Shipped
PRODUCT ENGINEERING CASE STUDY
Murmur
A private voice studio for Mac.
A native macOS product that turns private text and documents into natural speech while keeping generation on Apple Silicon after setup.
Visit the product
THE PRODUCT PROBLEM
Systems the app
had to coordinate.
Local speech systems expose model downloads, long-running generation, memory limits, voice selection, and file management. Murmur needed to turn that technical surface into a product people could install, understand, and use repeatedly.
KEY ENGINEERING DECISIONS
Choices that shaped
the implementation.
One product, several interfaces
The native studio, CLI, and MCP tools share the same underlying capability instead of becoming separate demos.
Direct Mac distribution
Licensing, signed releases, notarization, and updates are treated as core product systems.
DELIVERED
What the build includes.
Native SwiftUI production interfaceMultiple local speech enginesVoice cloning and Voice Design workflowsPDF and EPUB projectsBatch rendering, CLI, and MCP automationLicensing and direct update infrastructure
VERIFIABLE PRODUCT STATEPublic product availableRuns on Apple SiliconLocal generation after model setup
RELATED SERVICE
Building something with similar constraints?
NEW DESKTOP PRODUCTS, REBUILDS, AND RELEASE WORK
macOS App Development
DISCUSS YOUR APP
Need a similar native build?
Share the current product, target Apple platform, and workflows the app must support.
Discuss your app