OpenWispr
Star Download
OpenWispr vs Raycast Dictation

OpenWispr vs Raycast Dictation

Raycast Dictation is a feature of the Raycast launcher, not a standalone app — cloud-only, with no disclosed speech model, currently free during a beta with unannounced pricing after.

Download for macOS Get it on Google Play
OpenWispr/Compare/vs Raycast Dictation
The short version

A launcher add-on, not a dedicated dictation product.

Raycast Dictation ships inside the Raycast launcher on macOS and Windows (since June 2026), plus iOS through the Raycast Keyboard. It's cloud-only — Raycast does not publicly name its speech-to-text partner or model, and there's no user model choice. It's free during the current beta, with pricing after the beta unannounced. It has a 20-minute session cap and no Android build.

It does ship genuinely good ideas worth naming on their own merits: Auto Styling that picks a tone based on the app or website you're dictating into, Custom Vocabulary, Custom Instructions, single-tap modifier hotkeys, and team-shared Dictation Styles. Transcriptions and usage statistics are stored locally and never sent as telemetry — a real, specific claim, distinct from where the audio itself is processed.

OpenWispr
Raycast Dictation
Android app
Yes — on Google Play
No — macOS, Windows, and iOS (via Raycast Keyboard) only
On-device transcription
Always
No — cloud only, with no user model choice
Speech model / subprocessor disclosed
Yes — NVIDIA Parakeet-TDT (default) or Whisper, named and selectable
No — Raycast does not publicly name its STT partner or model
Works with no internet connection
Yes, fully
No — cloud dependency, connection lost means dictation stops
Price
Free, no tier, ever
Free during beta; post-beta pricing unannounced
Open source
Yes — MIT
No
Session length cap
None documented
20-minute cap per session
Custom vocabulary / auto-styling / team-shared styles
Personal dictionary planned, not shipped today
Yes — Custom Vocabulary, Auto Styling, Custom Instructions, team-shared Dictation Styles
Sourced from the competitor's own documentation, changelog, and public statements as cited below. Figures can change — see linked sources for the latest.
Where Raycast Dictation is ahead

Polish around the edges of the actual transcription

Auto Styling — automatically adapting tone based on the app or website you're dictating into — is a genuinely useful idea, as are Custom Vocabulary and Custom Instructions for steering output without a settings menu. Team-shared Dictation Styles is a distinctly B2B feature nothing else in this research offers. None of that requires knowing what model does the actual transcription, which is precisely the part Raycast doesn't disclose.

It's also worth crediting explicitly and separately: Raycast states transcriptions and usage statistics are stored locally and never sent as telemetry. That's a real, narrower privacy claim, distinct from — and not a substitute for — where the audio itself is processed, which is their cloud.

Where OpenWispr is ahead

It's Android, works offline, and tells you what model is running

Raycast Dictation has no Android build, no offline mode, and doesn't name its transcription model or subprocessor publicly — you're trusting a launcher plugin with your voice without knowing what processes it. OpenWispr runs on Android, works with no internet connection at all, and its model (Parakeet-TDT by default, or Whisper) is named, selectable, and checkable in open-source code.

Raycast's pricing is also an open question: free during the current beta, with no stated plan for after. OpenWispr's pricing has one state: free, permanently, because there's no cloud infrastructure cost driving a future price.

Questions

Good questions, honest answers.

Is Raycast Dictation free? +
Currently, during its beta. Pricing after the beta ends is unannounced. OpenWispr is free with no tier and no cloud infrastructure cost that would require a future price.
Does Raycast Dictation work offline? +
No — it's cloud-only, so a lost connection stops dictation. OpenWispr works fully offline after the initial model download.
What speech model does Raycast Dictation use? +
Not publicly disclosed. A widely repeated claim that it's "OpenAI GPT-5.4 mini" traces back to a competitor's marketing page (superwhisper's), not to Raycast — treat it as unverified. OpenWispr names its model: NVIDIA Parakeet-TDT by default, or Whisper via whisper.cpp, selectable.
Does Raycast Dictation have an Android app? +
No — it's a feature of the Raycast launcher on macOS and Windows, plus iOS via the Raycast Keyboard. There's no Android build. OpenWispr's Android app is live on Google Play.
Is Raycast Dictation open source? +
No. OpenWispr is MIT-licensed, with the full source on GitHub.
Also compare
OpenWispr vs FreeFlow (cloud by default)
Also compare
OpenWispr vs Wispr Flow (cloud-only)
Use case
Dictation software your IT team won't need to block

Start dictating privately.

Open source and entirely on your device. Android from Google Play, macOS from GitHub — no account, no sign-up, just a download.

Download for macOS Get it on Google Play
OpenWispr Voice to text, on your device.
GitHub Privacy Policy Features vs Wispr Flow vs superwhisper On-device Android dictation MIT © 2026