x-octo home Business judgment on AI products
中文

Real-revenue cases · AI

Murmur - Private, unlimited. Text-to-speech for Apple Silicon.

Real demand

A local text-to-speech app for Apple Silicon Macs that turns scripts into natural speech with cloned or custom-designed voices; one-time purchase, unlimited generation, audio never leaves the machine.

Stripe-verified
Monthly revenue
$1,202 / mo
Primary source
View original

01

Real-demand verdict

Real-demand verdict

Real demand

It replaces per-character or subscription cloud TTS services and the practice of uploading scripts to third-party servers; for Mac users who care about privacy and long-run cost, a one-time local tool is a reason to keep paying.

How it makes money

Customers pay a one-time purchase price and then generate unlimited audio; revenue comes from a Stripe-verified $1,202 per month.

What old behavior it replaces

Metered cloud text-to-speech services and the workflow of uploading scripts to external servers.

02

Where the first customers came from

Not mentioned in the source material

03

Tactics you can copy

  1. 01Make 'privacy plus unlimited one-time purchase' the core pitch, aimed directly at metered cloud TTS pricing.
  2. 02Support only Apple Silicon, trading local compute for server costs.
  3. 03Use voice cloning and custom voice design as the differentiator instead of generic reading.
Moving it to an AI business

For AI app builders: when a model can run locally, positioning 'data never leaves the device plus one-time purchase' as the alternative to metered cloud pricing avoids inference costs and attracts privacy-sensitive users.

04

Evidence and limits

Self-reported numbers are unaudited — treat them as leads, not facts
Stripe-verified

Revenue basis · Stripe-verified (TrustMRR)

What evidence is missing

The source does not state the purchase price, conversion rate, or user count, so the customer base behind the $1,202 monthly revenue is unknown.
No acquisition channels or cold-start details, so the playbook's replicability cannot be verified.
No retention or refund data, so the durability of a one-time-purchase model cannot be assessed.