Vocco / About

Changelog

What changed in each version of Vocco. Updates install from inside the app: when one is ready, the sidebar shows a restart button.

Updated October 28, 2026 · Written by the developer of Vocco

1.0.8 · October 28, 2026

• Dictation is more accurate. Vocco still recognizes each sentence while you speak, and now listens to the whole dictation once more when you let go. It keeps that result unless it looks wrong, such as cut short or turned into one language. In our test on 200 real recordings mixing Chinese and English, the error rate dropped from 10.4% to 9.8%. Dictations of several sentences take about a second longer to paste; a single sentence is as fast as before.
• New: Settings → Quick dictation → When text appears. Choose “As you speak” to have each sentence pasted at the cursor as soon as you pause, with the sentence you're still saying shown above the dot. “After you let go” stays the default because it's more accurate.

1.0.7 · October 28, 2026

There is no separate 1.0.6: the work planned for it shipped as part of 1.0.7.

• No sign-in needed: open Vocco and start talking. Every new Mac gets a 14-day trial with everything included, with no account and no card. The trial is tracked per Mac with a random install ID and a one-way hash of the Mac's hardware ID, never your name or email.
• When the trial ends, Vocco keeps working as a free version instead of locking: 30 minutes of dictation a week (Qwen3 included), and recordings and files up to 15 minutes each, without speaker labels. Vocco shows how many free minutes are left and tells you clearly when you reach a limit.
• Vocco Pro is now a one-time purchase that covers all your Macs and future updates. New purchases no longer offer a monthly plan; existing subscriptions keep working and can still be managed from Settings.
• Buying inside the app unlocks this Mac automatically once you've paid, even if paying takes a while. To use Vocco on another Mac, copy the activation code from Settings → Use on another Mac, or sign in there with Google or Apple using the email you paid with.
• Pasting an activation code now works without signing in. If a code can't be used (refunded, expired, or tied to a different account on this Mac), Vocco says why.
• A new "Try it" step in setup lets you hold right ⌘ and see your words appear before you start. Vocco asks for Speech Recognition up front when the built-in engine needs it, and permissions you denied earlier link straight to System Settings.
• New installs open at login (change it in Settings → General), so right ⌘ works right after a restart.
• When dictation produces no text, a short message next to the dot says why (microphone, permission or model) with a button to fix it.
• Recordings are written to disk as you record and autosaved every 30 seconds. If Vocco quits unexpectedly or the Mac loses power, the recording is recovered the next time you open Vocco. Recordings longer than 90 minutes keep their full audio.
• If nothing was recognized, the recording is still kept, so you can play it back and transcribe it again with another model.
• Select and delete many transcripts or dictations at once (⌘-click, ⇧-click, ⌘A), with a clear summary of what will be removed. The original files you transcribed are never deleted.
• File transcription shows each step (separating speakers, transcribing), explains why a file failed, and lets you retry it.
• Transcript titles show the actual date and time instead of "Today".
• Vocco now always uses the light appearance, which fixes unreadable menus and text fields in dark mode. The desktop pet has been removed.
• ⌘, opens Settings. Shortcuts that clash with system shortcuts are refused, with a one-click reset to right ⌘. Vocco asks before quitting during a recording and before signing out.
• Qwen3 ASR, the most accurate engine, now starts downloading in the background as soon as you first open Vocco. The built-in engine covers you until it's ready, then Vocco switches over by itself. Downloads resume after you quit, use several connections at once, and move to a faster source automatically when one is slow. Vocco doesn't start the download on Low Data Mode, a Personal Hotspot or a nearly full disk.

1.0.5 · October 27, 2026

• Hold right ⌘ to talk: hold, speak, release, and the text is pasted. A quick tap still starts and stops dictation.
• Speaker labels on long recordings no longer run your Mac out of memory. A leak in the speech runtime is fixed; an hour-long file with speaker labels now finishes in about 11 minutes on an M5 Pro.
• Fixed a rare crash when quitting while a transcription was still running.
• Sign-in now gives the app a secure session, so your subscription and license can only be managed from your own account.
• Qwen3 ASR no longer pastes stray text in other scripts (such as Thai or Korean) when you speak Chinese and English.
• With the Apple Speech engine, if macOS can't recognize your language on this Mac, Vocco now stops and tells you, instead of letting macOS send the audio to Apple.
• Japanese text is no longer converted to Chinese characters, and words from different languages are joined with the right spacing.
• First launch follows your system language, remembers the engine you picked even if you quit during the download, and suggests Qwen3 ASR when you are still on Apple Speech.
• Permission prompts are now in English and Simplified Chinese; Settings links to the open-source acknowledgements.
• The update button waits until you finish recording.

1.0.3 and 1.0.4 · August 17, 2026

• Two plans: $6.99 a month, or $69.99 once for a lifetime license. Pick either from the paywall.
• Manage or cancel a monthly plan from Settings.
• Your trial length now comes from your account, so reinstalling or switching Macs keeps the same trial.

1.0.2 · August 13, 2026

• Fixed the price shown in the paywall when the price service can't be reached.
• Wording updates across the app.

1.0.1 · August 12, 2026

• Speech models now download from an up-to-date source list, so a broken mirror can be fixed without waiting for an app update.

1.0.0 · July 22, 2026

First public release.

• Push-to-talk dictation in any app: hold right ⌘, speak, release, and the text is pasted at your cursor.
• Real-time transcription with punctuation and sentence breaks.
• Audio and video file transcription with speaker diarization and SRT export.
• Five on-device engines: Apple Speech, Qwen3 ASR, SenseVoice, Paraformer and Whisper.
• Updates install from inside the app.