AI assistants and search decision
VibeWhisper
A competent developer can reproduce the core push-to-talk + Whisper transcription flow, but delivering the polished, signed, cross-macos distribution and ongoing compatibility/support that make the paid product valuable requires extra time and distribution work—so building a narrow personal replacement is realistic, but matching the commercial product experience is non-trivial.
Visit website↗Built by Aleksandar Jovanovic, who ships 7 products in this index
Not priced
No pricing page we fetched carried a figure, so there is nothing to compare against. The build side is still real.
$100one-off56 h to build
$2.7/mo3 h/mo upkeep
No published price to break even against.
Open-source builds that already do this
Every project below is open source and already does this job today. Fork one, self-host it, or take the parts you need - the build prompt further down assumes an empty file, and this is the shortcut past that. Licences differ; check the one on each card before you ship. All VibeWhisper alternatives, with the arithmetic →
What a replacement has to do
- Capture microphone audio while a global push-to-talk key is held → send buffered audio to OpenAI Whisper API with user's key → receive transcript → insert text at cursor via macOS Accessibility API → discard buffer.
What it still won’t have
- Polished, cross-version macOS packaging and notarization workflow (codesign, notarize, installer)
- Polished UX edge-cases (shortcut conflicts, terminal focus quirks, accessibility permission flows)
- Ongoing compatibility and quick updates for new macOS releases across Intel and Apple Silicon
- Built, maintained commercial distribution channel, trial gating, and payment/key licensing flows
What remains hard
- Product polish and ongoing maintenance
First-year cost
No published price
VibeWhisper does not publish a price we could read, so there is nothing to compare against. What building costs is below.
Money you would actually spend
Time you would spend
—
What you would spend
What we assumed
The verdict above measures whether you could build it. This one is only about money.
Runnable build prompt
Build a native macOS menu-bar app in Swift (SwiftUI + Combine) that implements push-to-talk voice-to-text using OpenAI Whisper. Stack: Swift, AVFoundation/AVAudioEngine for capture, AppKit/Accessibility API for text injection, Keychain Services for secure key storage, URLSession for OpenAI API calls, and Xcode build tools for signing and notarization. Core features in scope: global configurable push-to-talk shortcut (register via CGEvent tap), low-latency audio capture buffered while key held, HTTP upload to OpenAI Whisper API using user's locally-stored API key, parsing and insertion of returned transcript at cursor via Accessibility API without using clipboard, basic menu-bar UI for entering API key and selecting shortcut, 15-minute trial timer, and app packaging instructions (codesign + notarize). Out of scope: server-side components, multi-user licensing server, offline Whisper model inference. Require: error handling for missing Accessibility permission, network failures, and API errors; unit tests for shortcut and API integration paths; end-to-end manual test plan for common editors and terminal apps.
How we checked
How the score was reached
- Partly verdict base52
- An open-source build was found+5
- 5 cited sources+3
- Price verified on pricing page+3
- Evidence score63
The base comes from the verdict. Everything under it is a check that either happened or did not, and each one is a fact frozen in this record rather than a judgement made at render time - so the same evidence always produces the same number.
How scoring works →Cited sources · 5
Every page the run actually retrieved.
- official productVibeWhisper — macOS Voice-to-Text | One-Time $19
- official pricingVibeWhisper Pricing — $19 One-Time, No Subscription
- official docsVoice-to-Text for Developers — VibeWhisper for macOS
- open sourceBeingpax/VoiceInk
- open sourceOpenWhispr/openwhispr
Integrity checks
What held up, and what did not.
