Dictanta Support

Open your meeting in Zoom, Teams, Google Meet, or Webex. Click the coral wave icon in your Mac's menu bar (top-right of the screen) or press ⇧⌘R. The picker will pre-select "This meeting" if a meeting app is detected. Click Record. The first time you do this, macOS will ask for Screen & System Audio Recording permission — click Allow. Dictanta only listens to the audio; it never records video or sees your screen.

Apple bundles system audio under the same permission category as screen recording. There is no separate "system audio only" permission. Dictanta uses ScreenCaptureKit in audio-only mode — we do not capture video frames or pixels from your screen. You can revoke the permission at any time in System Settings → Privacy & Security → Screen & System Audio Recording. If you prefer to skip this entirely, choose "Use microphone only" on the welcome screen — Dictanta will still record your voice without ever asking for screen permissions.

No. Dictanta does not join your call as a guest or bot, and it does not appear in the participant list of Zoom, Teams, Meet, or Webex. Recording happens locally on your Mac by capturing the audio that's already playing through your speakers — the same way QuickTime or screen-recording tools work. That said, depending on your jurisdiction (and your company policy), you may be legally or contractually required to inform other participants that the call is being recorded. Please check your local laws.

Audio files are stored on your Mac in the Dictanta app group container. They are never uploaded to our servers — we don't operate any server that receives audio. If you turn on iCloud sync in Settings, only your transcripts and meeting metadata are mirrored to your private iCloud — audio is explicitly excluded from sync. To delete a recording, right-click it in the Library and choose Delete, or use the Settings → Storage controls.

No. Summaries are generated by Apple's Foundation Models framework — the on-device large language model that ships with Apple Intelligence. The model runs entirely on your Mac. Dictanta does not call any third-party AI API. Your transcript is never sent to OpenAI, Anthropic, Google, or any other vendor. This requires Apple Intelligence (an Apple Silicon Mac, M1 or later, running macOS Tahoe). On Macs without Apple Intelligence the recording and transcription still work; the AI summary is unavailable.

Each summary bullet (TL;DR, action item, decision) is anchored to the exact moment in the audio it came from. Hover the bullet with your trackpad or mouse and the audio scrubs to that moment so you can hear the original sentence. Click the pin icon on a bullet to commit the playhead — useful if you want to leave the moment playing while you take notes elsewhere. Right-click any bullet for Play / Copy / Report options.

Three things to check, in order: (1) Make sure macOS Screen & System Audio Recording permission is granted for Dictanta — System Settings → Privacy & Security → Screen & System Audio Recording. After granting, you may need to quit and relaunch Dictanta. (2) Confirm the meeting app's audio is actually playing through your speakers, not a Bluetooth headset that bypasses the system mixer (some older Bluetooth profiles do this). (3) If the meeting hadn't started yet when you hit Record, system audio may have been silent — Dictanta still recorded, but there was nothing to capture.

Dictanta uses Apple's SpeechAnalyzer framework, which supports the languages Apple ships with that framework on macOS Tahoe — including English (multiple regions), Spanish, French, German, Italian, Portuguese, Japanese, Korean, and Mandarin Chinese, with more languages added by Apple over time. Language is auto-detected; if the result looks wrong you can pick a specific language for that recording.

iCloud sync moves transcripts only — the audio file stays on the Mac that recorded it. Make sure (1) you're signed in to the same iCloud account on both Macs, (2) iCloud Drive is enabled in macOS System Settings, and (3) you've toggled iCloud sync on in Dictanta's Settings on both machines. The first sync after enabling can take a few minutes depending on transcript volume.

Open any recording. In the toolbar, click the share button and choose Export as Markdown… or Export as JSON…. macOS will show a save panel — pick a location and click Save. Markdown is ideal for Notion, Obsidian, Apple Notes, or pasting into a doc. JSON includes timestamps and is useful if you want to import the transcript into another tool.

Dictanta uses Apple's standard subscription system. To cancel: System Settings → Apple Account → Subscriptions → Dictanta → Cancel Subscription, or open Dictanta → Settings → Subscription → Manage Subscription. Refund requests go through Apple at reportaproblem.apple.com — we don't have access to your billing information. Lifetime purchases don't auto-renew; you keep access forever and can refund through the same Apple flow within Apple's stated window.

Yes. Drag any audio file (.m4a, .mp3, .wav, .aiff, .caf, .flac) onto the Dictanta window, or use File → Import (⌘O). Dictanta will copy the file into your library and transcribe it on-device. The same audio-anchored summary works on imported files.

The on-device transcription and summarization frameworks (SpeechAnalyzer and Foundation Models) require an Apple Silicon Mac (M1 or later) running macOS Tahoe. Intel Macs do not have the Neural Engine these models depend on. Foundation Models specifically also requires Apple Intelligence to be enabled in System Settings.

Two steps: (1) In Dictanta → Library, select all recordings and delete them. This removes the audio from disk and the transcripts from local storage and iCloud (if sync is enabled). (2) Optional: drag Dictanta from /Applications to the Trash. There is no account to delete on our side because we never had one — Dictanta does not store any data about you on any server.