Talkify is a voice dictation app that lives in your Mac’s menu bar notch. Hold a designated key, speak, and text appears right where your cursor is. It runs Apple’s native SpeechAnalyzer engine — the one shipped with macOS 26 (Tahoe) — entirely on-device, so there’s no account signup, no network traffic, and no audio data leaving your machine. Built by indie developer Tornike Gomareli and released free and open source under MIT. Version 0.7.0, published on August 24, 2026, added “Dictate and Translate,” which converts what you say into another language on the spot — and that translation also runs fully on-device, via Apple’s native Translation framework.
Key Features
- Real-time insertion: The experience is that text is already inserted the moment you release the key. Benchmarks on the official site claim 123ms under identical conditions, faster than superwhisper, VoiceInk, Wispr Flow, and MacWhisper (these numbers were measured by Talkify itself on a MacBook Pro M4 Max running macOS 26.5; no third-party verification has been confirmed)
- Fully on-device processing: Speech recognition runs through Apple’s native SpeechAnalyzer/SpeechTranscriber engines, and read-aloud through AVSpeechSynthesizer — all of it stays inside your Mac. No network requests are made, and no audio files are written to disk
- Multilingual dictation & real-time translation: Supports 30 locales across 10 languages, including English, German, Spanish, French, Italian, Japanese, Korean, Portuguese, Cantonese, and Chinese. You can assign different languages to different keys — fn for one, right ⌥ for another — and with v0.7.0, holding right ⌘ while speaking translates what you say into the language set in Settings before inserting it. Translation from English into 11 languages requires no extra download; other languages fetch a model the first time
- Drag-and-drop file transcription: Drag an audio or video file onto the notch and transcription proceeds in the background, in parallel with dictation. When it’s done, the result appears in the notch — drag it to a folder or a text field, or click to copy
- Read Aloud and a customizable HUD: Have selected text read back in Apple’s native voices, optionally translating it first. The notch visuals (waveform, glow, and others), sounds, and shortcuts are all configurable, and lightweight usage stats such as word count and speaking pace are stored locally
Pricing
| Plan | Price | Key features |
|---|---|---|
| Free (open source) | $0 | Full access to dictation, multilingual translation, file transcription, Read Aloud, and everything else, with no restrictions |
Pricing reflects information as of August 2026. Talkify is released as open source under MIT, and the FAQ on the official site states plainly that there are “no accounts, no subscriptions, no trials.” There is no paid tier. Check the official site for the latest details.
Pros and Cons
✅ Pros
- Free and open source (MIT), with no account signup and no network traffic — reassuring on the privacy front
- Uses Apple’s native engine, so no additional model downloads are required, and the app itself is a lightweight 9.2MB download
- Per-key language and translation-target switching, handy for multilingual speakers and bilingual users
- Handles audio/video transcription and Read Aloud in the same single app
⚠️ Cons
- Requires macOS 26 (Tahoe) or later and Apple Silicon — Intel Macs and older versions of macOS are out
- The official FAQ states outright that recognition accuracy is on par with other tools, and that no cleanup pass is applied for typos or phrasing
- Text insertion for dictation and Read Aloud goes through the system clipboard, so if you run a clipboard history manager, the text may be recorded there briefly (around 0.5 seconds) — as noted in the official GitHub README
- A small project from a single indie developer, so support and roadmap continuity are uncertain
Comparison with Similar Services
| Criteria | Talkify | superwhisper | VoiceInk |
|---|---|---|---|
| Processing | On-device (Apple’s native SpeechAnalyzer) | Local speech recognition | Local execution (Parakeet V3) |
| License & price | Free, open source (MIT) | Check the official site | Open source |
| Response time in official benchmark | 123ms | 216ms | 324ms |
The speeds in this table come from a benchmark Talkify’s official site ran on its own (MacBook Pro M4 Max, macOS 26.5); no independent reproduction or verification has been confirmed. Treat them as reference figures.
Who It’s For
- People who want to get meeting notes and documents down quickly by voice instead of typing
- Bilingual users who want to speak in English and auto-translate into Japanese or another language while drafting
- Anyone who needs to batch-transcribe recorded audio and video files
- Privacy-minded Mac users who would rather not send audio data to the cloud
Summary
Talkify puts Apple’s native on-device speech recognition engine to work in a free, open-source package while claiming fast, responsive performance in its own benchmarks. It handles multilingual dictation and translation, file transcription, and read-aloud in a single app — though it’s worth keeping in mind that it requires macOS 26 and Apple Silicon, and that continuity is uncertain for a solo-developer project. Updates are coming steadily, with real-time translation added in August 2026, and if you’re looking for privacy-respecting voice input on your Mac, it’s worth a try.