OpenWhispr
OpenWhispr is an MIT-licensed dictation and meeting-notes app for macOS, Windows and Linux that calls itself the open-source alternative to Wispr Flow and Granola. It runs Whisper, NVIDIA Parakeet and other models locally, or cloud models with your own key, cleans up the text with a language model you choose, detects Zoom, Teams and FaceTime calls, and labels speakers on the device.
What does OpenWhispr do?
OpenWhispr's README calls it "the open-source and free alternative to WisprFlow and Granola". Press a hotkey, speak, and the text appears at your cursor. A second hotkey dictates in one language and pastes another.
It covers two jobs. The first is dictation, with local models (Whisper, NVIDIA Parakeet, Cohere Transcribe and others) or cloud models through your own API key, and a language-model cleanup step that can use GPT, Claude, Gemini, Groq, OpenRouter or a local model. The second is meetings: the README describes auto-detecting Zoom, Teams and FaceTime calls, live speaker labelling with voice fingerprints that carry across meetings, calendar integration with Google, Microsoft and Apple, and notes with folders and semantic search.
It also imports audio and video files including YouTube links, exposes a public API and an MCP server, and offers enterprise sign-on. Local Whisper is GPU-accelerated on Metal, CUDA and Vulkan.
Key facts
- Licence: MIT, which the README describes as free for personal and commercial use.
- Local models named: Orukeet, Whisper, NVIDIA Parakeet and Cohere Transcribe; cloud processing with your own API key is the alternative.
- Cleanup and assistant through GPT, Claude, Gemini, Groq, Tinfoil, OpenRouter or local models, with an optional screenshot as context.
- Meetings: auto-detection of Zoom, Teams and FaceTime calls, live speaker labels with voice fingerprints that carry across meetings, and Google, Microsoft or Apple calendar integration.
- Local Whisper is GPU-accelerated on Metal, CUDA and Vulkan; the app is Electron 41 built on whisper.cpp and sherpa-onnx.
- Notes with folders, semantic search and cloud sync; team spaces and web sharing; audio and video import including YouTube links; a public API and an MCP server; enterprise sign-on.
- Installers: macOS .dmg for Apple Silicon and Intel, Windows .exe, Linux .AppImage, .deb, .rpm and .tar.gz; from source it needs Node.js 24 or later.
- On Intel Macs, live speaker identification is unavailable and notes search falls back to keyword matching, per the README.
What does it replace, and where does it fall short?
For dictation it replaces Wispr Flow and Superwhisper, and it is the only app in that group that also runs on Linux with a cleanup step. For meetings it replaces Otter.ai and Fireflies.ai on the machine you are on, with the speaker labels done locally.
Where it falls short: it does not join a call as a bot on your behalf, the way Otter and Fireflies do. Notes sync and team spaces run through OpenWhispr's hosted cloud. Export formats are not listed in the README.
How does OpenWhispr run?
Installers for macOS (.dmg for both chips), Windows (.exe) and Linux (.AppImage, .deb, .rpm, .tar.gz). From source it needs Node.js 24 or later. It is an Electron app built on whisper.cpp and sherpa-onnx. The maintainers run OpenWhispr Cloud for sync and sharing, and team spaces are described as free for signed-in users; the README states no prices.
Who is OpenWhispr for?
A team that wants dictation and meeting notes in one app, on all three desktop platforms, with the option to keep every model local. Someone who only wants a dictation key with nothing else running should look at Handy.
What limits does the README state?
On Intel Macs, live speaker identification and voice fingerprinting are unavailable because the runtime they depend on stopped shipping Intel macOS builds, and notes search falls back to keyword matching.
Questions people ask
Is OpenWhispr open source?
Yes. The code is MIT, an OSI approved licence, and the README says it is free for personal and commercial use with no data collection and no telemetry. The maintainers also run OpenWhispr Cloud for sync and sharing, which the open-source app can use but does not require.
Does OpenWhispr keep my audio on my machine?
It can. Choose a local model such as Whisper or Parakeet and the audio is transcribed on the device; choose a cloud model and it is sent with your own key. Speaker labelling is done on device by the README's description. Notes sync and team spaces go through the maintainers' hosted service.
Which platforms does OpenWhispr run on?
macOS on both Apple Silicon and Intel, Windows, and Linux, with installers for each. One difference the README states: on Intel Macs, live speaker identification and voice fingerprinting are unavailable, because the runtime they depend on stopped shipping Intel macOS builds.
How does OpenWhispr compare with Wispr Flow and Otter?
For dictation it matches Wispr Flow's speak-and-paste with a cleanup step, and you choose the model behind it. For meetings it covers call detection and speaker labels on your own machine, where Otter sends a bot into the call. See the alternatives to Wispr Flow and alternatives to Otter.ai pages for the full comparisons.
Does OpenWhispr have an API?
Yes. The README lists a public API and an MCP server, so other tools and agents can read notes and transcripts. It also lists enterprise sign-on with SSO and SCIM and support for Amazon Bedrock or Azure OpenAI as model backends.
Sources
- OpenWhispr README and MIT LICENSE: github.com/OpenWhispr/openwhispr, read 2026-09-16.
Compared with the others
On the open-source alternatives to Wispr Flow page, OpenWhispr is the pick for cleanup by a model you choose, plus meetings. MIT. Local models or cloud models with your key, a cleanup step through GPT, Claude, Gemini, Groq or a local model, and call detection with speaker labels on device.
Also on that page: Handy for all three desktops, nothing else running, VoiceInk for a mac, with modes per application, FluidVoice for a mac with the newest local models.
On the open-source alternatives to Superwhisper page, OpenWhispr is the pick for cleanup by a model you choose, plus meetings. MIT. Local models or cloud models with your key, a cleanup step through GPT, Claude, Gemini, Groq or a local model, and call detection with speaker labels on device.
Also on that page: VoiceInk for a mac, with modes per application, Handy for all three desktops, nothing else running, FluidVoice for a mac with the newest local models.
On the open-source alternatives to Otter.ai page, OpenWhispr is the pick for call detection and speaker labels. MIT. Detects Zoom, Teams and FaceTime calls, labels speakers on device with voice fingerprints that carry across meetings, and reads Google, Microsoft or Apple calendars.
Also on that page: Meetily for live calls, entirely on your machine, Buzz for recordings, files and youtube links, Vibe for system audio and document exports.
On the open-source alternatives to Fireflies.ai page, OpenWhispr is the pick for call detection and speaker labels. MIT. Detects Zoom, Teams and FaceTime calls, labels speakers on device with voice fingerprints that carry across meetings, and reads Google, Microsoft or Apple calendars.
Also on that page: Meetily for live calls, entirely on your machine, Buzz for recordings, files and youtube links, Vibe for system audio and document exports.
More voice dictation apps
A free, offline push-to-talk dictation app for macOS, Windows and Linux that runs Whisper or NVIDIA Parakeet on your own machine and pastes the text at your cursor, under MIT.
A GPL-3.0 macOS dictation app that runs Nemotron, Parakeet, Cohere Transcribe, Apple Speech or Whisper on the device, with command and write modes, per-app prompts, and an optional closed local enhancement model.
A macOS dictation app under GPL-3.0 whose README names Superwhisper and Wispr Flow as the tools it replaces, with local models, per-app modes, a personal dictionary and an AI assistant mode.
A speech-to-text app inside the Epicenter monorepo under AGPL-3.0 that records, transcribes with a provider you choose, cloud or local, optionally polishes the text, and pastes it at the cursor.
An Apache-2.0 macOS dictation app that calls Gemini 3.5 Transcribe with your own key, writes audio to disk from the first millisecond for crash safety, and inserts polished text at your cursor without an account.
A free, open-source dictation app for macOS under Apache-2.0 that holds a shortcut, listens, and pastes the text where your cursor is, with Whisper, NVIDIA Parakeet or Apple on-device speech recognition.
Added September 16, 2026. Every claim above comes from the project's README, LICENSE or model card, read on September 16, 2026, or from the GitHub API on the date shown in the panel. Found an error? Write to reveneau@licheo.com and it is fixed in the next weekly pass. Repository: github.com/OpenWhispr/openwhispr.
Want OpenWhispr running on your own servers?
Reveneau sets it up on your hardware, connects it to your product, and builds the software around it.