EchoPad

Every call, on paper.

EchoPad records your calls and meetings on your Mac, tells the speakers apart and saves the transcript where you keep your notes. Free and open source.

Install EchoPad

Nothing leaves your Mac

  • Audio is recorded only between start and stop, into a folder on your Mac.
  • Transcription and speaker detection run on the Neural Engine.
  • The only network requests download the models from Hugging Face, once. After that EchoPad works offline.
  • No analytics, no crash reporting, no account. The source is open, so you can check.
Transcription settings: Parakeet TDT 0.6B v3, runs on the Apple Neural Engine, fully offline
EchoPad next to a typical meeting note-taker
EchoPadTypical note-taker
PriceFreeMonthly subscription
Audio is processedOn your MacOn remote servers
A bot joins the callNoOften
Transcripts go toAny folder you pickTheir web app
Source codeOpen, MITClosed

No bot joins the call

EchoPad hears what your Mac plays, not a meeting API. Your microphone and the call go on two tracks with one timeline, so your words are always yours and the other side is split into speakers you rename once.

Because nothing joins the call, nobody is told you are recording. Laws on recording differ by country and many require consent, so tell people yourself.

Recording settings: your microphone on one track, everything this Mac plays on the other

Read the privacy page

The EchoPad window: a Zoom call titled Weekly product sync, its transcript split between Lukasz, Anna and Marek

First, a call starts

Zoom, Teams, Google Meet in a browser, Slack, FaceTime, Webex, Discord, WhatsApp, Signal or Telegram takes the microphone, and EchoPad offers to record. It never starts on its own.

EchoPad

Zoom is using the microphone

Record and transcribe this conversation with EchoPad?

Record

One call, five files

One call, five files. Pick a format to see the file EchoPad writes for it.

2026-09-25 10.30 - Weekly product sync.md

---
title: "Weekly product sync"
date: "2026-09-25"
time: "10:30"
duration: "00:47:12"
app: "Zoom"
speakers: "Lukasz, Anna, Marek"
language: "en"
audio: "[[attachments/audio/2026-09-25 10.30 - Weekly product sync.m4a]]"
source: "EchoPad"
tags: "meeting"
---

# Weekly product sync

![[attachments/audio/2026-09-25 10.30 - Weekly product sync.m4a]]

[00:00] **Lukasz:** Morning both. Quick agenda today: the export redesign, the beta date, and what we heard from the interviews.

[00:07] **Anna:** I can start with interviews. Five of six people said they keep their notes in a folder they already sync, so they do not want another place for transcripts.

[00:19] **Marek:** That matches the support tickets. The most common question is still whether files can go straight into Obsidian.

[00:26] **Lukasz:** Good, then save locations stay the main feature. Each one gets its own folder, file name and formats.

[00:33] **Anna:** Can we add subtitles? Two people record talks and want an SRT next to the video.

[00:40] **Lukasz:** Already there. SRT and WebVTT are one toggle away, and they can live next to the Markdown.

It offers to record calls in

  • Zoom
  • Teams
  • Google Meet
  • Slack
  • FaceTime
  • Webex
  • Discord

Parakeet hears 25 languages

Detected automatically, or set in Settings. The code goes into the front matter.

  • Bulgarianbg
  • Croatianhr
  • Czechcs
  • Danishda
  • Dutchnl
  • Englishen
  • Estonianet
  • Finnishfi
  • Frenchfr
  • Germande
  • Greekel
  • Hungarianhu
  • Italianit
  • Latvianlv
  • Lithuanianlt
  • Maltesemt
  • Polishpl
  • Portuguesept
  • Romanianro
  • Russianru
  • Slovaksk
  • Sloveniansl
  • Spanishes
  • Swedishsv
  • Ukrainianuk

Built in the open

EchoPad is MIT licensed, and so are the two Swift packages it is made of. Use them in your own apps.

  • Conversations: today's Weekly product sync, 47 minutes on Zoom with three speakers

    47 min3 speakersZoom162 words

    25 Sep 202610:30

    Conversations, searchable

  • Save location editor: folder ~/Notes/Meetings, file name {date} {time} - {title}, subfolder {year}/{month}

    MarkdownTextSRTJSON

    ~/Notes/Meetings{year}/{month}

    Save locations, one per place

  • Recording settings: your microphone, the other side as system audio, permissions

    YouOther sideCore Audio2 tracks

    MicrophoneEverything on this Mac

    Recording settings

  • Transcription settings: language detected automatically, speakers identified, Parakeet TDT 0.6B v3 ready

    Parakeetv325 lang.Offline

    Detect automaticallyNeural Engine

    Transcription settings

Keep every conversation

Build it from source with three lines in Terminal. It needs macOS 26 and Xcode 26 or its Command Line Tools. Setup asks for two permissions, downloads the models and lets you pick where transcripts go.

Read the install guide

Terminal

~ % git clone https://github.com/pieralukasz/echopad.git
~ % cd echopad
echopad % scripts/install.sh