Getting started.

Before you start

• macOS 14.4 or later (that’s when Apple added the system-audio capture echo99 uses).
• Apple silicon and Intel Macs are both supported. On-device transcription is fastest on Apple silicon.
• An internet connection for the first model downloads. After that, echo99 works offline.

Finish setup

On first launch, echo99 opens a welcome window with a short checklist. Each row takes one click:

Welcome to echo99

Welcome to echo99

Records your microphone and system audio, then transcribes the call on-device.

Finish setup to start recording:

Microphone accessCapture your side of the call Grant

System-audio accessCapture the other side Grant

Transcription model (v3)Multilingual · change it later in Config 483 MB Download

Speaker modelTells "Them-1", "Them-2" apart 13.7 MB Download

✓ Storage location~/Documents/echo99 Choose…

✓ Recognize speakers across recordingsOn-device voiceprints

Continue

• Microphone access lets echo99 hear your side of the call. Click Grant and approve the macOS prompt.
• System-audio access captures the other side (what your Mac plays). Click Grant and approve.
• Transcription model (v3) is the multilingual speech-to-text model (25 languages), downloaded once to your Mac. Change it later in Config → Models.
• Speaker model tells the other voices apart, labeling them “Them-1”, “Them-2”, and so on.
• Storage location is where recordings are saved. Default is ~/Documents/echo99; click Choose… to change it.
• Recognize speakers across recordings is ticked by default. It lets echo99 keep an on-device voiceprint for each person you name, so it can label them in later calls. Voiceprints are biometric data and stay on your Mac. Untick it in the welcome window, or turn it off later in Config → Transcription.

The welcome window shows the size of each model download and its progress. The two models download from the internet once. After that, recording and transcription work with Wi-Fi off.

You’re ready

Welcome to echo99

Welcome to echo99

Records your microphone and system audio, then transcribes the call on-device.

Finish setup to start recording:

✓ Microphone accessCapture your side of the call Done

✓ System-audio accessCapture the other side Done

✓ Transcription model (v3)Multilingual · change it later in Config Done

✓ Speaker modelTells "Them-1", "Them-2" apart Done

✓ Storage location~/Documents/echo99 Choose…

✓ Recognize speakers across recordingsOn-device voiceprints

Continue

When every row is checked, click Continue. echo99 lives in your menu bar, with no dock icon and no window.

Automatic recording

Open the echo99 menu and turn on Auto-record calls once. The setting stays enabled across launches. When another Mac app uses your microphone for a few seconds, echo99 starts recording automatically and stops when the microphone is no longer in use.

Record manually

1. Click the echo99 icon in the menu bar → Record.
2. Have your call as usual. echo99 captures your mic and the system audio as separate tracks.
3. Click Stop Recording when you’re done. Transcription runs on your Mac automatically.
4. Click All recordings in the echo99 menu to open the Recordings window, then pick the call to read its transcript and play it back.

You can also start and stop recording from any app with the keyboard shortcut ⌃⌥R. Change it in Config → Shortcuts.

After your first call

• Read and play back. The Recordings window shows the transcript next to the audio. Play any line to hear what was actually said.
• Name the speakers. Type a name for “Them-1” and echo99 recognizes that voice in later calls. See speaker identification.
• Summarize. Open Summary to have a local language model write key points and action items. See meeting summaries.
• Export or re-transcribe. Export as Markdown… saves a speaker-labeled transcript; Re-transcribe… runs it again with a different model.
• Import a file. Import Recording… in the echo99 menu transcribes an existing audio or video file. See transcribing audio and video files.
• Watch it live. Turn on the experimental live transcript in Config → Experiments to follow the call while it records. See live transcription.

Where your recordings go

Each recording is its own folder under ~/Documents/echo99/, named by its start time (for example 2026-10-05_14-31-42). Change the location in Config → Storage.

• mic.wav: your microphone (labeled “Me”).
• system.wav: the other side (diarized into “Them-1”, “Them-2”).
• metadata.json: timestamps and device info.
• transcript.json / transcript.txt: the full transcript with speaker labels.

If a permission was denied

If you clicked “Don’t Allow” by mistake, echo99 can’t record until you re-enable it:

• Microphone: System Settings → Privacy & Security → Microphone → turn echo99 on.
• System audio: System Settings → Privacy & Security → Screen & System Audio Recording → turn echo99 on in the “System Audio Recording Only” list.

When a permission row in the welcome window turns red, its Open Settings button takes you to Privacy & Security. After you turn echo99 on, go back and click Grant again, or start a recording. If echo99 still reports the permission as off, quit and reopen it.