A MacWhisper alternative built for live meetings

Quick verdict

Side-by-side

During the call, or after it

MacWhisper watches a long list of apps for a starting call, Zoom and Teams and Webex and Slack and Discord and the browsers among them, and prompts you to record. When the meeting ends you stop the recording, it appears in the history sidebar, and you open it to transcribe. On a fast Mac that takes moments. It is still a second action on a queue you have to remember to clear, and the same page marks automatic meeting recording as beta: “there may be occasional issues or data loss.”

Silkwave Voice writes the transcript as the audio arrives, so you can read it while the call is still going. Pause, and the summary generates itself: key topics, action items, decisions. Nothing is queued and nothing is left to process.

MacWhisper Pro does have live captions with optional translation, from the microphone or system audio, though that is a separate feature from the meeting flow above. Neither app sends a bot into your call; both capture at the system level, which is how recording Zoom, Teams, and Meet without a bot works.

Models, and what you have to set up

MacWhisper's power comes from letting you pick the engine: Whisper from Tiny to Large v3, Parakeet on Apple silicon, distilled and custom builds, or a cloud provider you add. Since version 14 that list also includes Apple's own speech models on macOS 26, the same ones we use. That choice is why it reaches 100+ languages, and why you can trade speed for accuracy in a way we cannot. It is also a choice you have to make, and MacWhisper notes that its Medium and Large models want more than 8GB of RAM.

Silkwave Voice does not offer the choice. It uses Apple's built-in speech models, so setup is installing a language pack, five at most. Ten languages, one engine, nothing to benchmark. So on macOS 26 this is not an accuracy or a privacy difference at all. Both apps can run the same Apple models locally. It is a difference in how much you have to decide before you get a transcript.

The AI features work the same way. MacWhisper connects to OpenAI, Anthropic, Groq, xAI, DeepSeek, Azure and Ollama, and on macOS 26 to Apple's on-device Foundation Model, so summaries can stay on your Mac. The cloud providers need a key you supply and pay for. Silkwave Voice sends transcript text through Apple Intelligence to ChatGPT with no key and no separate bill. Neither sends your audio anywhere, a trade-off worth reading about on its own.

Pricing

Neither one is a subscription, so price decides less here than it usually does. MacWhisper's free tier is not a trial: drag-and-drop transcription, 100+ languages, dictation and meeting recording are in it, and free users can run the large Whisper models. Exporting is what Pro really unlocks, since a transcript made with a Pro-only model cannot be exported or copied. Pro is one payment of about €65, with lifetime updates.

Silkwave Voice is $24.99 once, after a 7-day trial, with no free tier behind it. Cheaper, and a much smaller app for the money.

One thing to know before you search: the App Store does not have MacWhisper. It has Whisper Transcription, a separate app from the same developer, priced separately and, per MacWhisper's own comparison of the two, without dictation, meeting detection or the CLI. On a Mac, the Gumroad version is the more capable one.

What Silkwave Voice does

• Notices when a call starts and offers to record it in one click
• Captures your microphone, the meeting's audio, or both, with no loopback driver to install
• Writes the transcript on-device as the call happens, in 10 languages
• Labels who said what, on-device and off until you turn it on
• Summarizes into key topics, action items, and decisions when you pause
• Answers follow-up questions about any recording in a side panel
• Transcribes audio and video files you already have, several at a time
• Searches every transcript you have ever recorded
• Records transcript-only, so no audio file is kept at all
• Runs from the menu bar without opening the main window

That is the whole list. Screenshots are on the product page, details in the documentation, and if you are weighing it against a cloud note taker instead, there is Silkwave Voice vs Granola.

Where MacWhisper is the better tool

It wins on languages. Over 100 against our ten. If your audio is Arabic, Hindi, Dutch, Polish or Turkish, this comparison ends there.

It wins on getting the text out. SRT and VTT subtitles, DOCX, PDF, CSV, Markdown, HTML. Silkwave Voice exports one .zip with the audio, a plain-text transcript and the summary, which is no use for subtitling video.

It wins on volume and automation. Batch queues, watch folders, a command line tool, and integrations with Notion, Obsidian, Zapier and webhooks, all of them Pro. Silkwave Voice imports several files at once, and that is the end of its automation story.

It wins on dictation. MacWhisper types into any text field on your Mac as a replacement for Apple's dictation. Silkwave Voice records; it does not type into other apps.

It wins on which Macs it runs on. macOS 14 or later, Intel included. Silkwave Voice needs macOS 26 for the on-device speech models it is built on, so if you are not on macOS 26 the comparison is over.

And on Pro it transcribes YouTube links, which Silkwave Voice does not. We built a separate free app for that, Silkwave Video.

Frequently asked questions

What is the difference between Silkwave Voice and MacWhisper?

MacWhisper is a file transcription app: you give it a recording and it produces a transcript, with Whisper, Parakeet or Apple's own models, 100+ languages, and subtitle and document exports. Silkwave Voice is built around live meetings: it detects a call, records it, and writes the transcript on-device while the call is happening, then summarizes it. It transcribes audio and video files you import as well, in 10 languages. Both run locally and both are a one-time purchase.

Does MacWhisper transcribe meetings in real time?

Not for meeting recordings. MacWhisper's documentation says that once you stop the recording it appears in the history sidebar, and you select it there to transcribe it. MacWhisper Pro does have a separate realtime captions feature that shows live subtitles from your microphone or system audio, with optional live translation.

Is MacWhisper free?

There is a genuinely free tier, not a trial. It covers drag-and-drop transcription, 100+ languages, system-wide dictation and meeting recording, and free users can run the large Whisper models, including Large v3 Turbo. The catch is on the way out: MacWhisper's release notes say exporting and copying a transcript are blocked when you transcribed it with a Pro-only model. Pro also adds Parakeet, speaker recognition, batch and watch folders, YouTube import, realtime captions, the CLI, and the Notion, Obsidian, Zapier and webhook integrations.

How much does MacWhisper Pro cost?

It is a one-time purchase in euros with lifetime updates. The macwhisper.com site lists €64 and the Gumroad checkout showed €65 when we last checked, and the developer notes the price changes over time, so confirm it at checkout. Bulk packs of 5, 10, 20 and 50 licenses are also sold, and students, journalists and non-profits can email for 25% off.

Is MacWhisper on the Mac App Store?

Not under that name. The App Store app from the same developer is called Whisper Transcription, it runs on Mac, iPhone and iPad, and it is a separate product with its own pricing: free to download, with a Pro subscription, a Pro Lifetime unlock, and an optional Assistant subscription for cloud transcription and AI. That lifetime unlock is priced very differently by country: $89.99 and $99.99 on the US store, but €53.99 and €59.99 in the eurozone, which undercuts the €65 Gumroad license. MacWhisper's own documentation says dictation, meeting detection, the CLI and full AI provider access are in the Gumroad version only, and that a Gumroad license key does not activate the App Store build. On iPhone and iPad there is no Pro unlock at all: the local models are free, and only the Assistant subscription costs anything.

Do I need an API key to summarize a meeting in Silkwave Voice?

No. Summaries and chat go through Apple Intelligence to ChatGPT, so there is no key to paste and no separate AI bill. You do need Apple Intelligence turned on with the ChatGPT extension enabled. Only transcript text is sent, never your audio, and the feature is off until you set it up.

Can Silkwave Voice export SRT or VTT subtitles?

No. Silkwave Voice exports a recording as a single .zip containing the audio, a plain-text transcript, and the summary as Markdown. If you need SRT, VTT, DOCX or PDF output, MacWhisper is the right tool.

How many languages does Silkwave Voice support?

Ten: Cantonese, Chinese, English, French, German, Italian, Japanese, Korean, Portuguese, and Spanish. MacWhisper supports over 100 through Whisper, so if your audio is in a language outside that list of ten, use MacWhisper.

Does Silkwave Voice run on macOS 15 or on an Intel Mac?

No. Silkwave Voice needs macOS 26 or later because it uses Apple's on-device speech models. MacWhisper runs on macOS 14 and later and works on Intel Macs, though its own page warns that performance on older Intel hardware can be poor.