Openwhispr
Voice-to-text dictation app with local (Nvidia Parakeet/Whisper) and cloud models (BYOK). Privacy-first and available cross-platform.
Last verified:
What is Openwhispr?
OpenWhispr is an open-source voice-to-text dictation app that turns your voice into text, notes, and actions from your desktop. Press a hotkey, speak, and your words appear at your cursor. It works across macOS, Windows, and Linux in any app that accepts text, including ChatGPT, Claude, Cursor, Slack, Google Docs, Gmail, Teams, and more.
The tool offers multiple transcription options: fully private offline transcription using local Whisper or NVIDIA Parakeet models where audio never leaves your device, cloud processing via OpenWhispr Cloud for instant transcription without API keys, or bring-your-own-key (BYOK) for OpenAI, Groq, and other cloud models. It includes AI cleanup features where you can give voice instructions like 'clean this up' or 'draft an email to Mike,' a custom dictionary that auto-learns names and jargon from your corrections, and support for 100+ languages with auto-detection.
OpenWhispr is designed for developers, writers, and teams who want fast dictation (3-5x faster than typing), privacy-first transcription, meeting transcription with automatic calendar integration, and AI agent mode. It features a local SQLite database for notes with optional cloud sync, semantic search, MCP integration for AI assistants, and a public REST API for programmatic access.
The app is free forever with local models and your own API keys, with optional Pro and Business plans for users who want zero setup, unlimited cloud transcription, and advanced features like agent mode and chat over your data.
Openwhispr pricing
Pricing model: Freemium
Free tier: $0 forever, includes OpenWhispr Cloud with 2,000 words/week cloud transcription, 5 hours of meeting recordings/month, unlimited local AI models, unlimited cloud transcription with your own API keys, 100+ languages, custom dictionary, zero data retention, and community support. Pro plan: $6.67/user/mo (or $8/month), includes all Free features plus unlimited cloud transcription, 20 hours of meeting recordings/month, sync across devices, personal API access, MCP integration in all apps, mobile app (iOS coming soon), and email support. Business plan: $16.67/user/mo, includes all Pro features plus unlimited meeting recordings, agent mode, chat over your data, and priority support. Enterprise plan: Custom pricing, includes all Business features plus managed API credits, cloud sync across devices, team features & admin, SSO & compliance, and dedicated support.
Openwhispr pros
- Open source with MIT license for free personal and commercial use
- Privacy-first: local processing means audio never leaves your device
- Zero data retention: audio not stored during cloud processing
- 3x faster than typing (~150 WPMS vs ~40 WPMS)
- Works offline with local Whisper and Parakeet models
- 100+ languages supported with auto-detection
- Cross-platform: macOS 12+, Windows 10+, Linux
- Custom dictionary that auto-learns from corrections
- Voice instructions like 'clean this up' for AI text editing
- Unlimited local AI models at no cost
- Bring your own API keys for cloud models
- Automatic meeting transcription with Google Calendar integration
- Speaker labels for meeting transcriptions via OpenAI Realtime API
- Full REST API with scoped permissions for programmatic access
- MCP integration for AI assistant connectivity
- Notes stored locally in SQLite with optional cloud backup
- No Input Monitoring permission required on macOS
- Multiple Whisper model options from Tiny (75MB) to Turbo (1.6GB)
Openwhispr cons
- Free tier limited to 2,000 words/week for cloud transcription
- Free tier only includes 5 hours of meeting recordings/month
- Local models require downloading (75MB to 1.6GB per model)
- Requires Microphone and Accessibility permissions on macOS
- Screen Recording permission needed for meeting audio capture
- Mobile app only coming soon for iOS, not yet available
- Cloud processing audio sent to third-party providers (OpenAI, Groq)
- Pro plan at $6.67/user/mo may be costly for some users
- No native Android app yet
- Self-hosting requires technical knowledge to build from source
Frequently asked questions about Openwhispr
Is OpenWhispr free?
Yes. OpenWhispr is open source and free to use. The free plan includes 2,000 words/week of cloud transcription, and local processing has no limits. Pro plans start at $6.67/user/mo for unlimited cloud transcription.
Which processing method should I use?
Use local processing for privacy and offline use. Use cloud processing for speed and convenience. Local models like Whisper Tiny run directly on your hardware with no internet needed after download, while cloud processing transcribes instantly without downloading models.
Can I use OpenWhispr commercially?
Yes. The MIT license allows commercial use, modification, and distribution. You can use it for personal and commercial projects without restrictions.
What languages are supported?
OpenWhispr supports 100+ languages for transcription including English, Spanish, French, German, Chinese, Japanese, Portuguese, Russian, Korean, Italian, Arabic, Hindi, Turkish, Dutch, Polish, and many more. Set your preferred language in settings or use auto-detect to switch mid-conversation.
Where are my notes stored?
Notes are stored locally in SQLite on your machine. Cloud sync is optional — when signed in to OpenWhispr Cloud, notes are backed up to the cloud but local storage remains the primary copy. The transcription history lives in a local database that we cannot access.
Is my data secure?
With local processing, audio never leaves your device. With cloud processing, audio is sent to the transcription provider (OpenAI, Groq, etc.). OpenWhispr has zero data retention — audio passes through for transcription and is never stored. No recordings, no logs, no copies. We never train on your data without explicit consent.
Does OpenWhispr need Input Monitoring on macOS?
No. OpenWhispr uses NSEvent monitors instead of CGEvent taps, so Input Monitoring permission is not required. Only Microphone and Accessibility permissions are needed on macOS, plus Screen Recording for meeting audio capture.
How does meeting transcription work?
Connect your Google Calendar, and OpenWhispr detects meetings automatically for Zoom, Teams, and FaceTime. Audio is transcribed in real time via the OpenAI Realtime API with speaker labels. The Free plan includes 5 hours of meeting recordings/month, Pro includes 20 hours, and Business includes unlimited recordings.
What local AI models are available?
OpenWhispr offers multiple Whisper models: Whisper Tiny (75MB, fastest), Whisper Base (142MB, fast), Whisper Small (466MB, balanced), Whisper Medium (1.5GB, accurate), and Whisper Turbo (1.6GB, best accuracy). NVIDIA Parakeet is also available. All models run directly on your hardware with no internet needed after download.
How do I get an API key for the OpenWhispr API?
Open the OpenWhispr desktop app, go to Integrations > API Keys, and create a key with the scopes you need. Your key starts with owk_live_ and is only shown once at creation. Available scopes include notes:read, notes:write, transcriptions:read, and usage:read. Keys have rate limits per minute and per day depending on your plan.