← all products / Voice Dictation
GhostType
- Category
- Voice Dictation
- Price
- $12/mo
- Verdict confidence
- medium
- Build time
- weekend
Verdict
Local speech-to-text models (e.g. Whisper) make the transcription itself free and buildable; the OS-level text-injection plumbing across every app is the fiddly part.
Core job
Hold a hotkey, speak, and have the transcribed (and lightly cleaned-up) text typed wherever the cursor is.
DIY replacement scope
A local hotkey listener piping audio to a speech-to-text model and injecting the result as keystrokes.
Requirements
- Node.js or a native helper for text injection
- a local or API speech-to-text model
Technical dependencies
- speech-to-text model (local or API)
- OS-level text injection
- global hotkey capture
What you lose
- polished cross-app formatting rules
- custom vocabulary tuning UI
- mobile app
Moat
- proprietary models
- execution polish
Why people still pay
Reliable cross-application text injection and per-app formatting rules take real polish to get right on every OS.
Pricing
$12/mo — Pro (flat)
Checked 2026-08-22 · confidence: medium · source ↗
Open-source alternatives
Build prompt
Build me a local push-to-talk dictation tool (fictional example — development fixture, not researched). Requirements: - A background app that listens for a global hotkey (hold to talk, release to transcribe). - On release, run the recorded audio through a local Whisper model (whisper.cpp or similar) and type the resulting text at the current cursor position via OS-level text injection. - A small settings window to pick the hotkey and the Whisper model size. - Out of scope: cloud transcription, custom vocabulary tuning UI, mobile app. - README: platform-specific setup for text injection permissions (macOS Accessibility, Windows, or Linux, whichever you target).
Recommended stack: Rust or Node.js native module, local Whisper model
Questions
What does GhostType do?
System-wide voice dictation with AI cleanup
How much does GhostType cost?
About $12/month (Pro), checked 2026-08-22.
Should I build a replacement for GhostType?
Local speech-to-text models (e.g. Whisper) make the transcription itself free and buildable; the OS-level text-injection plumbing across every app is the fiddly part.
What do I lose by replacing GhostType?
polished cross-app formatting rules; custom vocabulary tuning UI; mobile app.