It's why everything is a subscription these days: you make more money.
Consumers undercount the true total cost. And because X% of people will forget they're subscribed and keep paying forever.
If every month you had to either consent to recurring charge on your card or unsubscribe, I'm sure billions of revenue would evaporate overnight from people mass unsubscribing.
(I wish there was regulation that required companies to automatically pause monthly subscriptions if you haven't logged in to or used the service in any way for 3+ months. Though that would create some weird incentives)
That's why I'm building Knowii Voice AI. It's a fork of Handy (handy.computer) in which I'm going to add fun features, exploring different areas of what we can do with voice on a computer.
It's local first, privacy first, one-time payment. You buy it and get lifetime updates.
Currently available for Windows and very soon for MacOS and Linux. I'm working on Wayland/Hyprland support because I'm using Omarchy;-)
The same reason any software costs money even when you run it locally I suppose. Local software having a subscription instead of a single price seems to just be increasingly common these days.
I'd assume there are good free alternatives though. If not I'd have a non-zero motivation to build one, having dabbled enough with whisper and running several of my own distributed automatic transcription systems
Handy https://handy.computer/ is a free and open source voice transcription app. Works on macOS, Linux, and Windows, and you can pick which model. Everything runs locally. There is no API.
Its been years since I touched transcription professionally, but my memory is that they started off as a mechanical turk operation.
Lawyers usually would purchase transcription devices, and then either they would have a pool of transcribers (i remember installing foot pedals for forward/back playback operation) or pay a subscription to the manufacturer for mysterious likely offshore people to transcribe for them.
People have a hard time letting go of revenue, but I am betting most of the same people are still in business and want to pied piper consumers of transcription services to the same business model that now costs them pennies instead of wages.
This drive to turn one-time sales into repeating subscription fees has soured me even more on the concept of 'buying' (which really comes down to 'renting') software and makes it far less likely that I will ever pay to use these products. I'll go rather great lengths to avoid any software which comes with a price tag not even so much for the actual price - if I count time invested in finding, building, installing and maintaining free (as in freedom as well as beer) software alternatives to paid proprietary products it probably comes out even or more 'expensive' - but to avoid the whole licence/renew/upsell/'gold-silver-platinum-Pro-whatever' dance and the accompanying lock-in. It is a bit like how the onslaught of online advertising has turned me from being somewhat tolerant towards banner advertisements into a rabid content blocker who makes sure not a single piece of advertising ever gets to pollute my eyes or ears. Squeeze too hard and you'll find your hands empty before you realise it.
This is great! I’ve been diving deep into local models that can run on this kind of hardware. Been building this exact same thing, but for complete recordings of meetings and such because, why not? I can even run a low-end model with ollama to refine and summaries the transcription. Even combining with smaller embedding models for a modern, semantic search. It has surprised me how well this works, and how fast it actually is locally.
Hopefully we will see even more locally run AI models in the future with a complete package.
I'm deaf, so I test a lot of speech to text and transcription apps from an accessibility point of view.
My answer to "why have a monthly subscription" would be that you need capabilities that Whisper doesn't handle well, like real-time transcription in noisy environments.
That's not the niche you're targeting here, though. :)
My experience is that Whisper - not being built for real time speech to text - isn't as good at it as other tools are. You can hack something together by stacking together progressively more audio frames to feed to Whisper to give it context, but IME, you're going to get better results from a model that's designed for real-time STT in the first place, or by using a service like Azure Speech to Text which has excellent noise resilience... but which is also an ongoing cost which would justify a subscription. Real-time Whisper also devours your battery quickly.
That said - while I've had very good experiences with Parakeet in MacWhisper, I'm curious if you evaluated Apple's SpeechAnalyzer APIs at all. It's unfortunately limited macOS/iOS/iPadOS 26+ since it's a new API, but it's on device, has comparable quality of results to Whisper Large v3 Turbo and Parakeet, and seems to be better on battery usage.
21 comments
[ 531 ms ] story [ 404 ms ] threadI got frustrated paying monthly for something that could run on my Mac, so I built Lucid Voice:
- 100% offline (Nvidia Parakeet + Llama)
- $20 one-time (mainly to cover Apple's notarization costs)
- Runs on surprisingly low-end hardware (M1 base models work fine)
- No cloud, no data collection
Open to feedback on anything - pricing, the tech stack, or if this should just be free: https://lucidvoice.app
[0] https://goodsnooze.gumroad.com/l/macwhisper
Consumers undercount the true total cost. And because X% of people will forget they're subscribed and keep paying forever.
If every month you had to either consent to recurring charge on your card or unsubscribe, I'm sure billions of revenue would evaporate overnight from people mass unsubscribing.
(I wish there was regulation that required companies to automatically pause monthly subscriptions if you haven't logged in to or used the service in any way for 3+ months. Though that would create some weird incentives)
It's local first, privacy first, one-time payment. You buy it and get lifetime updates.
Currently available for Windows and very soon for MacOS and Linux. I'm working on Wayland/Hyprland support because I'm using Omarchy;-)
https://voice-ai.knowii.net
The price tag is $30/YEAR. The current MRR is about $700 and I'm paying $7/mo for Groq Whisper Turbo.
These apps really don't have any reason to be so pricey, it's all just margin.
1. https://whispertype.com/
I'd assume there are good free alternatives though. If not I'd have a non-zero motivation to build one, having dabbled enough with whisper and running several of my own distributed automatic transcription systems
I use the app constantly, all day long.
Lawyers usually would purchase transcription devices, and then either they would have a pool of transcribers (i remember installing foot pedals for forward/back playback operation) or pay a subscription to the manufacturer for mysterious likely offshore people to transcribe for them.
People have a hard time letting go of revenue, but I am betting most of the same people are still in business and want to pied piper consumers of transcription services to the same business model that now costs them pennies instead of wages.
I’ve found that Whispering [https://github.com/EpicenterHQ/epicenter/tree/main/apps/whis...] plus Groq pay-as-you-go is a great combination. Not quite free, but cheap enough that it isn’t a consideration.
https://support.apple.com/guide/iphone/view-a-transcription-...
Hopefully we will see even more locally run AI models in the future with a complete package.
Spokenly is free (one time fee of $0) and does the same (and even more)
My answer to "why have a monthly subscription" would be that you need capabilities that Whisper doesn't handle well, like real-time transcription in noisy environments.
That's not the niche you're targeting here, though. :)
My experience is that Whisper - not being built for real time speech to text - isn't as good at it as other tools are. You can hack something together by stacking together progressively more audio frames to feed to Whisper to give it context, but IME, you're going to get better results from a model that's designed for real-time STT in the first place, or by using a service like Azure Speech to Text which has excellent noise resilience... but which is also an ongoing cost which would justify a subscription. Real-time Whisper also devours your battery quickly.
That said - while I've had very good experiences with Parakeet in MacWhisper, I'm curious if you evaluated Apple's SpeechAnalyzer APIs at all. It's unfortunately limited macOS/iOS/iPadOS 26+ since it's a new API, but it's on device, has comparable quality of results to Whisper Large v3 Turbo and Parakeet, and seems to be better on battery usage.