Velora

What does Velora do?

Velora describes itself as "private voice typing for Mac, free, open source, and fully on-device." You hold the push-to-talk key (Right Option by default), speak, and release. Velora removes filler words, adds punctuation, follows spoken line breaks and keeps the original transcript in History. The text is pasted into messages, email, notes, documents and other text fields, and the final result stays on the clipboard even if the current app ignores automatic paste.

A separate shortcut (Ctrl+Shift+A) runs an action. The README gives examples like "search YouTube for the match highlights" and "message Priya on Slack that I'm running late." Velora opens the app and does the task a few steps at a time, reading what the screen says to decide the next step, and stops rather than guess when the window it expects is not in front. Say "draft" to stop with the message written but unsent.

Key facts

• Licence: MIT on the Mac app.
• Platforms: Apple Silicon Mac (M1 or newer) with macOS 14 or later. Intel Macs are not supported.
• Speech recognition and writing cleanup run on the Mac; audio and transcripts are not sent to a transcription service.
• Shortcuts: Right Option for dictation (default), Ctrl+Shift+A for agentic actions.
• Features: filler removal, automatic punctuation, spoken line breaks, a Personal Dictionary for names and terms, selected-text editing from a spoken description, meeting mic plus computer audio capture after confirmation.
• When dictation starts, Velora can pause supported playback in Apple Music or Spotify and resume only the playback it paused.
• Install: brew install --cask sushilk1991/tap/velora, or the signed macOS release from GitHub.
• The repository also contains an iPhone companion. The README says the iPhone app is alpha and is not stable at all; it is not on the App Store, has no TestFlight build and should not be depended on.

What does it replace, and where does it fall short?

Velora replaces Wispr Flow and Superwhisper for a Mac user who wants dictation plus a separate spoken-command shortcut, under MIT, with no account and no cloud. It appears on the open-source alternatives to Wispr Flow and Superwhisper pages.

Where it falls short: Apple Silicon only. The README does not state which speech model Velora uses, which language model handles cleanup, download sizes, latency or accuracy on any benchmark. The iPhone companion is explicitly alpha and unstable. The agentic action shortcut depends on reading what the screen says, so an unexpected window layout can stop a run.

How does Velora run?

Install with Homebrew: brew install --cask sushilk1991/tap/velora. Or download the signed Mac app from the releases page. On first launch, grant Microphone, Input Monitoring and Accessibility access. Accessibility is what lets Velora paste into other apps and run spoken actions.

The agentic actions mode works a few steps at a time, reads what the screen actually says, adjusts when a shortcut did not land, and stops rather than guess when the window it expects is not in front. Meeting capture records microphone and computer audio only after you confirm, and creates a local transcript and notes.

Who is Velora for?

A Mac user on Apple Silicon who wants local dictation and also wants a spoken-command shortcut that can open apps and carry out small tasks, under MIT. For dictation only, Handy covers the same push-and-paste shape on macOS, Windows and Linux. For a hosted product with team features, Wispr Flow or Superwhisper is the paid tool.

What limits does the README state?

• Apple Silicon Mac with macOS 14 or later; nothing else is supported.
• The iPhone companion app is alpha and is not stable at all. Not on the App Store, no TestFlight build, far less testing than the Mac app.
• The README does not state which speech model, language model, latency or accuracy figures Velora uses.
• The agentic action shortcut depends on reading what the screen shows; Velora stops rather than guess when the expected window is not there.

Questions people ask

Is Velora open source?

Yes. The LICENSE file at the repository root is MIT, an OSI-approved licence. The README says Velora is free, MIT-licensed, with no subscription and no account. The maintainer notes that the Mac app is the Velora product; the iPhone companion is alpha and should not be relied on.

Does Velora work offline?

Yes. The README says speech recognition and writing cleanup run on the Mac; audio and transcripts are not sent to a transcription service. There is no account to sign into and no telemetry. The README does not state which models Velora uses or how large they are, so check the first-run download when installing.

What does the agentic action shortcut do?

Hold Ctrl+Shift+A and say what you want done. Velora opens the app and acts a few steps at a time, reading what the screen says before the next step. The README gives examples such as "search YouTube for the match highlights" and "message Priya on Slack that I'm running late." Say "draft" to stop with the message written but unsent. Velora stops rather than guess when the expected window is not in front.

How does Velora compare with Wispr Flow?

Wispr Flow is a paid product that does dictation only, hosted. Velora is free and MIT, on-device by default, and adds a spoken-command shortcut that opens apps and reads the screen to carry out small tasks. The README's own table says Velora is on-device while Superwhisper and Wispr Flow are "mostly cloud or hybrid," and the agentic action feature is listed as present in Velora and not in those paid tools. The open-source alternatives to Wispr Flow page sets Velora beside Handy, OLIV and others.

Can I use the iPhone app?

The README says the iPhone app is alpha and not stable at all. There is no App Store build and no TestFlight. The Mac app is the Velora product. If you build the iPhone app from source, expect crashes and changes that break without warning between commits.

Sources

• Velora README and LICENSE: github.com/sushilk1991/velora, read 2026-11-02.