OpenWispr

Key features

• Fully offline Whisper and Parakeet transcription Fully offline Whisper and Parakeet transcription: speech is converted on the device by downloadable local models, from tiny fast Whisper sizes up to Turbo plus NVIDIA Parakeet, so dictation audio never leaves the machine and no internet is needed.[1][2]

• Global-hotkey dictation into any app Global-hotkey dictation into any app: a system-wide hotkey starts recording over any application, and the transcript appears at the cursor with automatic pasting into editors, mail, chat, and browsers.[1][2]

• AI cleanup and voice instructions AI cleanup and voice instructions: commands such as clean this up or draft an email to Mike apply light editing or rewriting to raw dictation, and a dedicated voice-assistant hotkey sends speech straight to the AI assistant as a command.[1][2]

• Custom dictionary with auto-learn Custom dictionary with auto-learn: names, jargon, and medical, legal, or domain-specific terms can be added, and the dictionary learns from corrections to keep specialized dictation spelled correctly.[1][2]

• Meeting transcription with diarization Meeting transcription with speaker diarization: Zoom, Teams, and FaceTime calls are auto-detected with live speaker labelling, voice fingerprinting, and Google, Microsoft, or Apple calendar integration, plus searchable notes.[1][2]

• AI agent across model providers AI agent across model providers: a named voice assistant talks to hosted models or local models, with optional screen context, streaming answers into a floating panel or pasting at the cursor.[1][2]

• Notes, sync, and team spaces Notes, sync, and team spaces: notes support folders, semantic search, cloud sync, and AI actions, while signed-in users get web sharing links and collaborative team spaces with roles and invitations.[1][2]

• Audio import and translation dictation Audio import and translation dictation: existing audio and video files can be dragged in, batch-uploaded, or transcribed from a pasted URL, and a dedicated hotkey dictates in one language and pastes the text in another.[1]

Who will love it

• Privacy-conscious dictators of sensitive material: on-device transcription with zero data retention means spoken words stay off hosted services unless you explicitly choose cloud processing.[1][2]

• Cost-sensitive users who want unlimited dictation free: the free tier never meters local transcription, so heavy dictators pay nothing while keeping 100-plus languages and the custom dictionary.[1][2]

• Meeting-heavy note takers on desktop: automatic call detection, speaker-labelled transcripts, calendar integration, and searchable notes compress a recording, transcription, and notes workflow into one app.[1][2]

Who should avoid it

• Users wanting unlimited cloud transcription for free: the free plan caps hosted transcription at 2,000 words per week and meeting recordings at 5 hours per month, so cloud-heavy use requires Pro or Business.[1]

• Intel Mac owners wanting every meeting feature: live speaker identification and voice fingerprinting are unavailable on Intel Macs because the required runtime stopped shipping macOS x86_64 builds, though recording and transcription still work.[1]

• Mobile-first dictators: the core product is a desktop app, and the iPhone and iPad app arrives with the paid Pro tier rather than as the free entry point.[1][2]

Alternatives and compatibility