v0.6.0-beta8

Released 2026-06-16 · Download .dmg (34 MB)

Three improvements driven by real meeting feedback. Two cover edge cases that quietly degrade transcripts, and one keeps polishing the voice-sample feature shipped in beta7.

A heads-up when you’re about to record without headphones

When you record a meeting with the laptop’s built-in speaker instead of headphones, the speaker plays the other person’s voice out loud into the room — and the built-in mic picks it up. Because the mic stream is always labeled “You,” those bleed-through utterances end up tagged as you in the transcript. You’ll see “You” next to things you didn’t say.

Beta8 detects that configuration at the “Before recording” dialog and surfaces a soft callout: “Headphones recommended — your output is the built-in speaker. Other people’s voices will play in the room and the mic will hear them too — they may show up labeled as ‘You.’” Not a hard gate — recording still works — just a heads-up so you can plug in if you have headphones nearby.

Plug headphones in mid-recording? Now you get a clean stop instead of a silent freeze

Before: switching audio devices in the middle of a recording (plugging in AirPods halfway through, swapping to a USB mic, etc.) made the live transcript stop updating with no warning. The framework had auto-stopped our audio engine and we weren’t catching it. You’d sit there wondering if the app crashed.

Now: Minutiva listens for that event, gracefully ends the recording with everything captured up to the change saved, and surfaces a clear message: “Audio device changed mid-recording. The meeting was saved with audio captured up to that point — start a new recording to continue.”

Voice samples skip leading silence automatically

Beta7 picked the longest segment for each speaker, which fixed the “played a 200ms back-channel” problem. Beta8 goes further: the player now scans the first few seconds of the chosen segment for actual voiced audio and jumps straight to it, skipping any quiet stretch at the start of the segment.

So when you click a speaker’s play button, you hear the voice immediately instead of a second or two of room tone first. The progress bar accounts for the trimmed silence, so the visual still matches what you hear.

Carried over from beta7

Nothing changed in

Auto-update

Existing beta channel testers get this within 24 hours. Or trigger now: Minutiva Private → Check for Updates…