Alex Godbehere

Trace - No-frills offline meeting transcripts with context

A macOS menu-bar app that turns any conversation into a clean markdown transcript, with a local speech model running entirely on-device. One global shortcut brings up a small bar at the bottom of your screen. It captures your mic and the system audio as separate tracks, labels who said what, and lets you flag key moments mid-call that sit inline at the right timestamp. No bot joins the call, nothing leaves your Mac, no account, no subscription.

Add a comment

Replies

Best
Alex Godbehere
Hey Product Hunt 👋 I'm the maker. I built Trace, a Mac app that records and transcribes your meetings entirely on-device. I built Trace for myself. I'd been using MacWhisper, but there was enough fiddling before each call that I'd forget to start it and walk out of an hour-long meeting with nothing written down. So the things I cared about most were that it's quick to activate and stays out of the way. You hit one global shortcut and a small bar appears at the bottom of your screen (there's also a keystroke to hide it entirely if you'd rather not see it while it's working). As it records your meeting you can flag anything important, with an optional note, as a "key moment". I built that because the "wait, that part matters" thought never survives to the end of a call. The key moments sit inline in the markdown at the right timestamp, so any AI you later paste the output into can see what mattered. It records your mic and the system audio as two separate tracks, then runs the system side through on-device diarization so you get who said what rather than one blended wall of text. Your mic lines are labelled as you, and the other voices come out as Speaker 1, Speaker 2 and so on (speaker renaming is planned for a future update). The transcription, and the live recap while you're recording, both run entirely on your machine, and you can choose between two engines. One is fast, using NVIDIA's Parakeet-TDT (Core ML, on the Neural Engine). The other is more accurate, using WhisperKit and Whisper large-v3-turbo, and holds up better on accents, jargon and quiet rooms. Both download their weights once and then run locally. There are no meeting bots to join the call. It just captures the audio your Mac is already playing out loud. Trace doesn't do any of the summarising itself, it just hands you clean markdown. On privacy, the app is sandboxed and your audio never leaves the Mac. The only network call it has to make is on first run, downloading the speech and speaker models (around 500MB) from Hugging Face, and after that you can go fully offline and it keeps working. The one other thing that ever touches the network is an optional Google Calendar connection for auto-naming sessions, which is read-only and stays off until you turn it on. You don't need to create an account, and we don't collect any telemetry. Trace is £9.99 once on the App Store, no subscription, and it needs macOS 14 or later on Apple Silicon since the models run on the Neural Engine. If you've used Granola, Otter or Fathom, the difference is they put a bot in your call or do the work in the cloud, sometimes both, and I wanted neither. I've been using it every day for months now and it's genuinely fixed the problem for me. Feedback very welcome, roasts included, and I'd love to hear whether it helps anyone else the way it helped me.
Denitsa Pencheva-Valtchanova

Congrats on the launch, this looks super useful (and the fact that files never leave the computer + pricing is a one-time payment, not a subscription, is veryy refreshing). One question on the key moments feature: how easy is it to export or push those specific flagged spots into another note-taking tool like Notion after a call?

Alex Godbehere

@denitsapenchevavaltchanova super easy! Every recording ends up on disk as a .md and a .json file so you can grep/jq/manually extract it. Key Moments live under their own key in the .json file for exactly this reason.

Here's both files from a transcript that I captured from a YouTube video

# 
Massive Surprises From Starship Flight 12!
_2026-05-26 13:10 · 1m 36s_

**Speaker 1** [00:00] Loads and loads of great diagrams and animations at the start of the stream showing the hot stage system difference, the payload area and the capabilities there. We had the alterations with the grid fin systems, Starlink magic like this, and even a neat announcement by Chun Wang from the FRAM2 mission saying that he intends to fly the first human interplanetary flyby.

> ⭐ **[00:22] Key moment:** First interplanetary flyby announced!

**Speaker 1** [00:22] That's a little far down the line of course, but the excitement with the future was dialed right up. Yes, there was so much crammed in here, but today we are focused on the flight. We've spent months following all the construction and assembly work leading to this moment. The giant flame trench, the upgraded water deluge system, all of it had finally led to this first real test. I've gotta say, to me this felt a bit more like watching flight 1 again and with everything untested together for the first full flight, it felt like anything could happen. Here we were though approaching the expected T-Zero and there was nothing left to do but cross those fingers and nervously hold my breath to see if the detonation suppression system deluge and ignition with fire up in sequence and lead to the most powerful rocket in history launching into the sky. Ignition.

> ⭐ **[01:10] Key moment:** Liftoff!

**Speaker 1** [01:20] Wow, was that incredible or what? This almost felt surreal and immediately one thing became obvious.

> ⭐ **[01:25] Key moment:** Cleared the tower

**Speaker 1** [01:25] All those Raptor 3s are belting out an unbelievably
**Speaker 1** [01:30] violent wake behind. Let's just rewind and take a closer look at this in more detail.
{
  "keyMoments" : [
    {
      "at" : 22.19532799720764,
      "note" : "First interplanetary flyby announced!"
    },
    {
      "at" : 70.8701651096344,
      "note" : "Liftoff!"
    },
    {
      "at" : 85.09528303146362,
      "note" : "Cleared the tower"
    }
  ],
  "segments" : [
    {
      "end" : 69.05999755859375,
      "speaker" : "Speaker 1",
      "start" : 0.05999999865889549,
      "text" : "Loads and loads of great diagrams and animations at the start of the stream showing the hot stage system difference, the payload area and the capabilities there. We had the alterations with the grid fin systems, Starlink magic like this, and even a neat announcement by Chun Wang from the FRAM2 mission saying that he intends to fly the first human interplanetary flyby. That's a little far down the line of course, but the excitement with the future was dialed right up. Yes, there was so much crammed in here, but today we are focused on the flight. We've spent months following all the construction and assembly work leading to this moment. The giant flame trench, the upgraded water deluge system, all of it had finally led to this first real test. I've gotta say, to me this felt a bit more like watching flight 1 again and with everything untested together for the first full flight, it felt like anything could happen. Here we were though approaching the expected T-Zero and there was nothing left to do but cross those fingers and nervously hold my breath to see if the detonation suppression system deluge and ignition with fire up in sequence and lead to the most powerful rocket in history launching into the sky. Ignition."
    },
    {
      "end" : 90.07999420166016,
      "speaker" : "Speaker 1",
      "start" : 80.87999725341797,
      "text" : "Wow, was that incredible or what? This almost felt surreal and immediately one thing became obvious. All those Raptor 3s are belting out an unbelievably"
    },
    {
      "end" : 94.97999572753906,
      "speaker" : "Speaker 2",
      "start" : 90.07999420166016,
      "text" : "violent wake behind. Let's just rewind and take a closer look at this in more detail."
    }
  ],
  "session" : {
    "endedAt" : "2026-05-26T12:12:34Z",
    "startedAt" : "2026-05-26T12:10:58Z"
  }
}

Hope that helps!

Denitsa Pencheva-Valtchanova

@alexgodbehere that's super neat!

Vasyl

Really nice idea 👍
How does the local speech model perform in noisy meetings or with multiple speakers overlapping?

Alex Godbehere

@workout097_collab Good question. The honest answer is I haven't stress-tested either engine in really noisy or heavy-overlap calls. Both Parakeet and Whisper large-v3-turbo are widely used elsewhere so there's a lot of public testing on them, but I can't tell you firsthand how they do in a busy cafe.

A decent mic makes a real difference (and helps with cross-talk from other participants, because your voice is recorded on its own track). The workflow I built Trace around is pasting the transcript into an LLM for a summary afterwards, which is surprisingly good at filling in context even when the transcript is a bit rough in places.

Eric Lagarda

On-device speech is the hard part to get right, and the fact that nothing leaves the Mac is exactly why people end up trusting it. Splitting mic and system audio into separate tracks is a detail most tools skip. I've been through the same on-device vs cloud tradeoff doing local receipt scanning, accuracy is the thing that keeps you up at night.

How's the local model holding up next to the cloud ones?

Alex Godbehere

@ericlagarda Local vs cloud is definitely a trade-off. In my experience the on-device models aren't anywhere near as good as the frontier cloud ones, but as it stands that's the price we need to pay for privacy. I keep an eye on new releases and hope to pull in better models as they become available, but I also want Trace to stay lightweight and not chew through battery or memory, so it's a balance.

I built this around pasting the transcript into a local LLM afterwards for the summary, and many local models are very good at filling in context and figuring out what was meant, even when the raw transcript is a bit rough (I've found Gemma V4 to be pretty good at this).

Samir Asadov

On-device transcription with no cloud round-trip is huge for sensitive calls — the use case I see most often is exactly the kind of conversation people don't want flowing through a SaaS pipeline. I host the ModeLoop Podcast on financial modeling and have wrestled with the privacy-vs-quality trade-off for years; the on-device speech models have finally crossed the threshold where local-first is a real choice, not a compromise. Following.