🇬🇧 English  ·  🇩🇪 Deutsch

CallNotes

You take a call on your Mac — CallNotes records both sides as separate tracks, transcribes locally with Whisper, separates speakers, and drops a finished, AI-summarized note wherever you want it. Fully automatic, from the menu bar.

_{🪟 On Windows? There's an experimental sibling: callnotes-windows}

---

Why this exists

Every call-recording tool either needs a virtual audio driver (BlackHole/Loopback), a visible meeting bot, or a cloud subscription. CallNotes needs none of that:

• Core Audio process taps (macOS 14.2+) capture the system audio of just the call app — your caller lands on its own track, background music doesn't.
• Your microphone is recorded in parallel — two separate tracks means perfect speaker attribution for 1:1 calls, no AI guessing needed.
• For conference calls, local speaker diarization (sherpa-onnx) splits the remote mix into "Speaker 1..N" — you match names via short audio snippets and a dropdown.
• Transcription runs on-device — whisper.cpp (Metal) or NVIDIA Parakeet TDT v3 (the fastest option, 25 EU languages, no repetition loops) — or via Groq API if you prefer cloud speed. Your choice, one toggle.
• Bring your own AI for summaries: Claude Code (default), any OpenAI-compatible API (OpenAI, Groq, OpenRouter), fully local via Ollama — or none at all.

What you get after hanging up

A finished Markdown note, ~1 minute later:

• Summary, discussed points, commitments & to-dos, open questions (Claude, optional — pick the sections you want, including a follow-up email draft)
• Dialog transcript with speakers ("Me: … / Caller: …"), timestamps included
• Stereo audio archive (left = you, right = them)
• Delivered to your notes folder (Obsidian-friendly), plus optionally Apple Notes, Nextcloud, Notion, an external drive mirror, and an ntfy push

How it works

The menu bar app

Everything lives in the menu bar (phone icon):

• Live view during a call — two animated level tracks (you + caller), call timer, and a popup to type in participant names while you still remember them
• Processing status after hangup (transcribing → speaker detection → AI summary)
• Speaker assignment for conferences: play a voice snippet per detected speaker, pick the name from a dropdown (Claude pre-suggests names it heard in the call)
• Recent calls, storage locations (incl. external drive), API keys, integrations
• First-run setup wizard, a built-in help section, and an ⓘ explainer next to every setting — you never have to guess what a field does
• Live mic monitor — the panel shows in real time which app is currently using your microphone, even one that's not in your call list (e.g. a browser). One click on "Always record this app" adds it, so browser calls get captured from then on.
• English & German UI — follows your system language automatically

Install

1. Dependencies (Homebrew + Xcode Command Line Tools):

brew install whisper-cpp ffmpeg

2. Clone & install:

git clone https://github.com/michaelczesun/callnotes && cd callnotes
./install.sh

This builds CallNotes.app into /Applications (the settings & status app — find it in your Applications folder or as the phone icon in the menu bar) and the invisible recording helper calltap.app into ~/Applications (don't move it — the recording permissions are tied to it). It also sets up the speaker models (~35 MB), creates ~/.config/callnotes/config.json and starts both background services.

3. Whisper model (once, ~550 MB — on 8 GB machines ggml-small.bin is fine too):

mkdir -p ~/models && curl -L -o ~/models/ggml-large-v3-turbo-q5_0.bin \
  https://huggingface.co/ggerganov/whisper.cpp/resolve/main/ggml-large-v3-turbo-q5_0.bin

4. Permissions — do this once. Click the phone icon in the menu bar; the setup wizard opens on first run (later: Settings → Restart initial setup). In the Permissions step, click "Request & check permissions now" — the two macOS dialogs appear (Microphone + System Audio Recording for "calltap"), allow both, and the button confirms with a green check. Note: calltap only shows up in System Settings' permission lists after this first request — that's macOS behavior, not a missing entry.

5. Test call (>20 s) — about a minute after hanging up, the note appears in your notes folder. Curious? tail -f ~/CallNotes/log/process.log.

Updating later: cd callnotes && git pull && ./install.sh — settings and notes stay untouched.

Uninstall

./uninstall.sh          # removes daemons + apps; keeps notes, config & models
./uninstall.sh --purge  # also removes working data, config, venv and models

Your notes folder is never touched.

Supported call apps

FaceTime, iPhone calls via Continuity, WhatsApp, Zoom, Teams, Signal, Telegram, Discord — anything that uses the microphone. The allowlist lives in the config; find any app's bundle ID with calltap procs --watch during a call (unknown apps are also logged automatically).

Configuration

Everything is in ~/.config/callnotes/config.json — or just use the menu bar settings. Highlights:

Key	What it does
apps	bundle IDs that trigger recording
tapScope	app = record only the call app family (default), global = all system audio
transcriber / groqApiKey	local (whisper.cpp), parakeet (fastest, ./install.sh --with-parakeet) or groq (cloud)
summarizer (+ summarizerUrl/Model/ApiKey)	claude (Claude Code CLI), openai (any OpenAI-compatible API incl. Ollama/Groq/OpenRouter) or off
noteSections	which sections Claude writes: summary, discussed, todos, follow-up email
destinations	extra delivery: Apple Notes, Nextcloud (WebDAV), Notion
notesDir / audioDir / mirrorDir	where notes, audio and the external-drive mirror go
diarize / diarizeThreshold	multi-speaker detection on/off, clustering threshold
speakerSelf / context	your name in transcripts + one line of context for better summaries

CLI

calltap procs [--watch]     # which app is using the mic right now?
calltap record --out DIR    # manual two-track recording (Ctrl-C stops)
bash process-call.sh DIR    # (re)process a recording
bash callnotes-sync.sh      # mirror notes + audio to the external drive

Troubleshooting

• The permission button shows green but calltap is missing from the "Screen & System Audio Recording" list: that's fine — on newer macOS versions tap-based apps may appear under "System Audio Recording Only" instead, or not be listed at all despite being granted. The source of truth is the daemon log: grep "Self-Test" ~/CallNotes/log/callwatch.log — "Systemaudio-Tap ok" means recording works.
• The system-audio dialog never appears (mic dialog works) — and other tap-based tools fail on your Mac too: that's a machine-level issue, not this app. Most common on company-managed Macs: an MDM/PPPC profile can block screen/system-audio capture prompts system-wide (check System Settings → General → Device Management, or profiles list in Terminal). On unmanaged Macs, a stuck permission database can be reset with tccutil reset ScreenCapture at.dasgeht.calltap, then request again via the wizard button. The daemon log names the exact error code: grep "Tap verweigert" ~/CallNotes/log/callwatch.log.
• System track is silent (-91 dB): the tap API returns silence, not an error, when the System Audio Recording permission is missing. Check System Settings → Privacy & Security → Screen & System Audio Recording → calltap.
• No permission dialog ever appeared: calltap must run as an app bundle via launchd (a bare CLI binary has no prompt identity). ./install.sh handles this.
• Recording never starts: tail -f ~/CallNotes/log/callwatch.log — if it logs "Mikro aktiv bei nicht gelisteter App", add that bundle ID to apps.
• Caller audio missing in Electron apps (WhatsApp/Discord/Teams): audio often plays in a helper process; tapScope: "app" covers the whole app family. If it still misses, set "global".
• Switched audio output mid-call (connected AirPods): the running capture can go silent — switch before the call, or accept the gap.
• Failed jobs land in ~/CallNotes/failed/ with raw audio; reprocess with bash process-call.sh <dir>.

FAQ

Does it work with browser calls (Google Meet, Teams web)?

Yes. Meet/Teams-web/Whereby run inside your browser, so out of the box they're not in the call list. When a browser call is active, the panel's live mic monitor shows it (“Chrome is using your microphone”) with an Always record this app button — one click and browser calls are captured from then on. Caveat: it then taps the whole browser's audio, so if you have music playing in another tab it can bleed into the recording — mute other tabs during calls.

Does it run on Windows or Linux?

This repo is macOS-only (14.2+) by design — the driverless two-track capture is built on Core Audio process taps. But there is an experimental Windows sibling built on WASAPI process-loopback capture with the same pipeline and config format: callnotes-windows — CI-compiled, looking for testers. Linux (PipeWire per-app streams) would work the same way — open an issue if you'd use one, PRs welcome.

Why no App Store / signed binary?

It's a git clone && ./install.sh tool for now. Everything builds locally in seconds; nothing is downloaded except the Whisper and diarization models.

I opened Teams/WhatsApp and nothing happened?

That's expected: the popup and the recording start when a call is actually running (the app is using your microphone) — not when you merely open the app. Call someone and watch the menu bar icon change. Also: CallNotes is a menu bar app (phone icon, top right) — since 1.2.1, double-clicking it in Finder opens the panel as a window, and the panel tells you whether the recorder service is running (green dot).

Intel or Apple Silicon?

Both. The requirement is macOS 14.2+, not the chip — Core Audio process taps exist on Intel Macs too. On Intel, Whisper transcribes without Metal acceleration (slower); consider the ggml-small model or the Groq transcriber there.

Which call apps are supported?

Anything that uses the microphone: FaceTime, iPhone calls via Continuity, WhatsApp, Zoom, Teams, Signal, Telegram, Discord — see Supported call apps above for adding your own.

Privacy & legal

Everything runs locally by default (Whisper on-device); only the summary — if enabled — goes to the Claude API, and transcription goes to Groq only if you opt in. Tell people you're recording. Laws differ by country (e.g., secretly sharing recordings is criminal in Austria, secretly making them is criminal in Germany). You are responsible for lawful use.

License

PolyForm Noncommercial 1.0.0 — free for personal and noncommercial use. Selling this software or using it commercially is not permitted.

---

_{🇩🇪 Diese Seite auf Deutsch}