Record any conversation. Keep every word on your device.
Record on your Mac or iPhone and read a transcript with speaker labels. Apple Watch recordings transfer to the paired iPhone for transcription. Transcription runs locally, and you choose whether to keep audio for playback. AI is optional; cloud choices ask before receiving selected text.
Keep the context with the conversation.
A short note, an agenda, or a decision from last week's meeting can help you pick up the thread. Add the context you need, check the source quotes, and keep what helps.
Add context to a note.
After analysis, choose Add context. Your original note stays separate. Read the source quotes, then Keep, Edit, or Discard the suggestion.
Bring the agenda.
Attach a readable PDF, Markdown, or text file to one recording or a recurring meeting series. Your original file stays untouched.
Ask across your library.
Questions can use saved recordings, notes, flags, and attached text. Open source quotes to check the passage, meeting, or PDF page.
Recall earlier decisions.
Optional Meeting context shows relevant quotes from earlier saved meetings while you record, alongside earlier open action items. It uses local search and stays out of the way until useful.
Add context and answers use the AI option you choose. Cloud AI asks before sending selected transcript, note, or attached-file text. Recorded audio is never sent to AI.
The difference is where your audio goes.
Recording, transcription, and speaker labels run locally. Keep notes, find important moments, and export readable files. Choose whether to retain audio for playback and whether to use optional AI. Recorded audio is never sent to an AI provider.
The trade you're making, stated plainly: there is no sync between your devices and no cloud library. Your Mac and your iPhone each keep their own recordings, and you move a library with a backup file you make and restore yourself. The app is English only.
One app. Your Mac, your iPhone, your wrist.
One purchase covers all three apps. Record calls on Mac, conversations on iPhone, and quick notes on your wrist. Each device keeps its own library.
Mac The full studio
Your Mac sees the most of any conversation.
- Records your mic and your Mac's system audio in separate lanes, and shows which are actually running
- Captures both sides of a Zoom, Meet, or Teams call, with no bot in the room
- A system-wide shortcut, ⌃⌥⌘R, starts and stops from any app
- Notices when a meeting app comes to the front and offers to record
- Printing, drag-and-drop, and a recording in its own window
iPhone In your pocket
The same studio, wherever the conversation is.
- Records the room through the mic, transcribing live
- Shortcuts, Siri, the Action button, a widget, and a Control Center control
- A Live Activity on the Lock Screen and in the Dynamic Island while you record
- Set the speaker count first for sharper labels
- Share-sheet export — Word, PDF, Markdown, HTML, SRT, VTT, Obsidian, Apple Notes
Watch On your wrist
For the conversations your phone isn't in your hand for.
- Records on the watch, then hands off to your iPhone to transcribe
- Elapsed time stays legible on the always-on display
- A complication showing the running timer and anything still waiting to send
- Every transfer listed with its status, the reason it failed, and retry
- Speaker count set with the crown, and remembered
Mac & iPhone Read, review, and export
Watch recordings join your iPhone library after transfer.
- Whisper transcription that runs entirely on-device
- Recognizes returning voices and labels them — on-device
- With kept audio, tap a line to play it; edit notes and flags long after
- Search across transcripts, notes, flags, and attached text
- Optional summaries — on-device, your key, or Synopsule Pro
- Backup and restore, which is also how a library moves between devices
On the Mac
The full studio





- 01
Library, transcript, and summary — together.
Pick a recording on the left, read the labeled transcript in the middle, and the optional OpenAI cloud summary shown here sits beside it.
- 02
A directory of every voice.
Speakers follow people across your recordings — and Synopsule recognizes returning voices, labeling them automatically. Rename someone once and it updates everywhere — transcript, search, and exports.
- 03
Your mic and the call, in separate lanes.
Core Audio taps record your microphone and system audio independently — both sides of a Zoom, with no bot in the room.
- 04
Exports that fit where the work lives.
Word, PDF, Markdown, HTML, SRT, and VTT — or one tap into Obsidian or Apple Notes — with a live preview, and speakers and timestamps kept intact.
- 05
Local by default — and you can see it.
Recording and transcription stay local. Choose whether to keep audio, and control automatic summaries in Settings.
- 06
Find the moment across everything.
Search transcripts, notes, flags, and attached text. Use
speaker:filters to find a voice, or turn on optional on-device search to find related wording.
On iPhone
In your pocket





- 01
Every recording in one place.
Starred, drafts, and the full archive — the people in each one shown at a glance.
- 02
See who said what.
Colored speakers and timestamps make the transcript easy to follow. With kept audio, tap a line to play that moment.
- 03
Notes and flags, as you record — or long after.
Jot time-anchored notes and drop a flag while you talk, then tap either later to seek right back to the moment. You can also add, rename, or delete them days later while reading. The live draft transcript runs underneath, with transcription on the phone.
- 04
Names that stick — and recognize.
Name a voice once and Synopsule recognizes that person in your next recordings on this phone, labeling them automatically — all on-device. Your Mac works the same way.
- 05
Find any moment.
Find words in transcripts, notes, flags, and attached files. Recordings you never summarized are searchable too, with
speaker:filters to find a voice. - 06
Choose how AI works.
AI setup explains summaries, answers, and note context together. Choose supported on-device AI, your own key, or Pro. Cloud AI asks before receiving selected text.
Built for how conversations actually go.
Long calls. People talking over each other. A hallway chat you wish you'd written down. Synopsule turns all of it into a searchable transcript, with speaker labels, notes, and exports you can use elsewhere.
Whisper, running on your hardware.
A draft transcript appears as you talk, with final processing after you stop. A Whisper model ships inside the app, so recording and transcription work offline from the first launch. Download a larger model, up to 1.5 GB, if you want another balance of speed and accuracy.
- Transcribes out of the box — a model ships in the app
- Keep several model sizes, switch instantly with no redownload
- Works offline, always · built for hour-long sessions
- Apple silicon Mac on macOS 15+; iPhone on iOS 17+
Capture the call, or the room.
On the Mac, Synopsule records your microphone and whatever is playing through your speakers in separate lanes, using Apple's Core Audio taps. No screen recording, no virtual driver, no bot in the meeting. On iPhone, it records the room through the mic: point it at the conversation and hit record. And the recording screen shows what is actually being captured right now — not what you once set in preferences — so a permission you never granted or a lane that failed to start says so before you rely on it.
- Mac: both sides of Zoom, Meet, or Teams, in separate lanes
- Mac: choose which input device, and start from anywhere with ⌃⌥⌘R
- iPhone and Watch: the whole room through one mic
- The capture state you see is the one you've got — degraded starts and errors are surfaced
It learns who's who.
When a recording finalizes, on-device diarization separates the voices — and Synopsule recognizes people it has heard before, labeling them automatically. Name someone once and the next time they speak, they're tagged for you. Correct a wrong guess and it learns. A bundled neural voice model does it all on-device, so the voiceprints never leave.
- Recognizes returning voices — matched lines show a “Recognized” chip
- Correct a speaker mid-recording; the fix survives finalization
- On-device neural voice model — voiceprints never leave
- Rename once; it propagates everywhere
Play it back, synced to the line.
Turn on Keep audio in Settings to save new recordings for playback. Tap a line to jump to that moment, change the speed, or skip fifteen seconds. The waveform comes from the saved audio. Your transcript, notes, flags, and find-and-replace still work when audio is deleted.
- Variable speed, ±15s skip, keyboard transport
- A waveform computed from the recording itself
- Find that steps through matches — plus find and replace
- Notes and flags you can add, edit, or delete at any time
Summaries when you ask — or the moment it's ready.
Pick a template — one-on-one, standup, sales call, interview, lecture, clinical SOAP, legal, or user research — and turn an hour of talk into a structured, speaker-aware recap with owners, tasks, and the questions still open. Press Summarize, or switch on automatic summaries and it happens as soon as analysis finishes. Re-run it and the action items you've already ticked stay ticked; switch templates and the recap you had is kept, not thrown away. Run it through your own key, on-device via Ollama or Apple Intelligence so nothing leaves, or Synopsule Pro.
- Automatic after analysis, if you turn it on — and off is the default
- Re-runs preserve your checked action items; template switching is non-destructive
- Your OpenAI, Anthropic, or Gemini key; on-device via Ollama or Apple Intelligence; or Synopsule Pro
- Export to Word, PDF, Markdown, HTML, SRT, VTT, the audio itself — or straight into Obsidian and Apple Notes
Summary
Locked the launch date for October. Engineering will stage the export redesign to fit; marketing starts long-lead comms immediately.
Action items
- Jordan — send updated risk register tonight
- Priya — draft customer comms by Thursday
- Maya — confirm launch date with leadership
Start when the conversation does.
Enable calendar awareness to see reminders for the calendars you choose, then decide when to record. Calendar events are read locally and never changed. Chosen meeting details can stay with the recording. Notifications tell you when a transcript is ready; on Mac, meeting-app prompts and ⌃⌥⌘R help you start. On iPhone, use Siri, Shortcuts, the Action button, a widget, or a Control Center control, with a Live Activity while recording.
- Calendar-aware prompts that name the recording after the meeting
- Notifications when a transcript is ready, or a Watch recording lands
- Siri, Shortcuts, the Action button, a widget, a Control Center control, a Live Activity
- Import an audio file you already have and transcribe it like a recording
Record it? You can use the meeting's name.
52 minutes · 4 speakers · transcribed on this Mac
Hallway note · 3 minutes · transcribing now
A library that survives a year of meetings.
Sort it, filter it, tag it, select several at once, move through it from the keyboard. Recurring meetings collect into a series so a weekly standup reads as one thread. Every open action item across every recording gathers into one list, grouped by who owns it or when it's due. And Backup writes a plain ZIP of your recordings — one you can open with any unzip tool — which restores by merging rather than overwriting, so it's also the way to carry a library from a Mac to an iPhone.
- Sort, filter, tags, multi-select, keyboard navigation
- Recurring meetings grouped as a series
- Action items across every meeting, by owner or by due date
- Backup and restore that merges — and moves a library between devices
Action items · by owner
- Jordan — send the updated risk register (Launch review)
- Priya — draft customer comms (Weekly sync, 3 of 3)
- Maya — confirm the date with leadership (Launch review)
Privacy you can diagram.
Recording and transcription run on-device. Cloud AI asks before receiving selected text, and recorded audio is never sent to an AI provider.
Mic captured to a temporary file. On Mac, system audio too. A Watch recording travels to your iPhone, not to a server.
Whisper transcribes on your Mac or iPhone. Speakers are labeled on-device.
New audio is deleted after successful processing unless you choose to keep it. Your transcript and notes remain.
Calendar events are read locally. Chosen meeting details can be saved with the recording. Notifications stay local.
Selected text goes straight to the provider you choose. Your key is stored in this device’s Keychain and used for those requests.
Selected text routes through Synopsule to OpenAI for the AI task you choose. No provider key to manage.
Your library does not sync. Move recordings with a backup file you create and restore yourself. Read the full Privacy Policy.
Readable before it's pretty.
A transcript is a reading tool. Synopsule follows your device's text, motion, and accessibility settings so you can read it comfortably.
Text that scales
Every font on iPhone and Apple Watch used to be a fixed size.
- Dynamic Type honoured throughout, up to the accessibility sizes
- Layouts that reflow instead of clipping when text grows
- Touch targets of at least 44pt
Colour that isn't the only signal
Speaker colour is how you tell voices apart while skimming, so it was solved rather than chosen.
- Eight speaker colours re-derived to stay apart under protanopia, deuteranopia, and tritanopia
- Distinct shapes per speaker when Differentiate Without Color is on
- Text held to AA contrast on every surface
Motion and VoiceOver
The ambient movement is decoration; the app works with all of it switched off.
- Reduce Motion honoured on Mac, iPhone, and Apple Watch
- VoiceOver gets real state, traits, and actions — not bare labels
- Progress, playback position, and speed are spoken as values, not left to a moving bar
Your conversations, ready for AI workflows.
Export readable Markdown with headings, speaker labels, and timestamps. On Mac, compatible AI apps can search your saved text and, with your opt-in, operate recording, editing, analysis, Library workflows and exports through MCP. Audio is never returned to the assistant.
# Launch Review
May 27, 2026 · 21m · 2 speakers · on-device
## Transcript
**Avery** · 00:00
Let's confirm the source separation before we publish the recap.
**Jordan** · 00:14
Speaker names update across every matching segment as soon
as the user confirms them.
## Summary
- Launch recap aligned; source separation confirmed.
- Exports carry clean timestamps and speaker labels.
- Structure a model reads on sight — headings, speakers, and timecodes, no copy-paste cleanup.
- Speaker labels and timestamps ride along, so the AI knows who said what, and when.
- It's just text. No lock-in and no special importer — paste it wherever you think.
Let your assistant work through the whole conversation.
Connect Claude Desktop, Claude Code or Codex in Settings → Connect to AI Apps. Search saved transcripts immediately. Switch on Allow AI app control and keep the Mac app open to record, edit speakers and notes, analyze, manage authored Library items, play audio locally, import files, create backups and export documents. You still approve permissions, AI destinations, file choices and deletion. Audio, voiceprints and credentials are never returned; the AI client controls what happens to text it reads. Read the setup guide.
- Claude DesktopOne-click install from Synopsule
- CodexApp, CLI, and IDE — one config
- Claude CodeOne terminal command
- ChatGPTDrag a Markdown export in, or tunnel
- Capture and analyzeStart, pause, stop, transcribe and generate recaps
- Review and editCorrect speakers and text, add notes, complete actions
- Search and organizeRetrieve sources and manage authored Library items
- Play, import and exportLocal playback, native file choices, documents and backups
- 38 discoverable toolsFive transcript tools and 33 opt-in app-control tools
Try: "Start a meeting recording." Then: "Rename the speakers, add a note at two minutes and export a Word document."
Or ask your meetings without leaving Synopsule.
Ask questions across saved recordings, notes, flags, and attached text. Inspect the source quotes in the answer; meetings you never summarized are searchable too. Open a quote to check its source, revisit recent questions, or cancel a running answer. Ask uses the AI option you configure; Pro supplies cloud AI without a provider key to manage.
One purchase. Every device. Forever.
Synopsule is a one-time $4.99 on the App Store — the whole app, on your Mac, your iPhone, and your Apple Watch. AI is optional. Choose on-device AI where supported, your own provider key, or Synopsule Pro.
Synopsule
App Store · Mac, iPhone & Watch- On-device Whisper
- Mic + system audio on Mac
- Recording on iPhone and Watch
- Speaker labeling & rename
- Playback, notes, and flags
- Search transcripts, notes & files
- Every export format
- Calendar prompts & notifications
- Shortcuts, Siri, widget, Live Activity
- Backup and restore
- PDF, Markdown & text attachments
- Optional local Meeting context
Synopsule Pro
Optional · covers every device- Summarizes a whole hour in one pass, so a long meeting keeps its detail
- All nine templates return their own sections — Clinical SOAP, Legal, Sales Call, and the rest
- Ask your meetings answers from long excerpts across several transcripts, with citations
- No API key and no provider account to manage — billing runs through the App Store
Selected text goes through Synopsule to OpenAI for cloud AI. Recorded audio is never sent. On-device AI keeps that text here. Subscriptions renew until cancelled, and you cancel in App Store → Subscriptions.
Get Synopsule, then add Pro in the appAI choices without a Pro subscription
Use a provider key, local Ollama on Mac, or Apple Intelligence on supported devices. Provider pricing and hardware requirements depend on your choice.
Selected text goes straight to your chosen provider. You pay that provider for usage; your key is stored in this device’s Keychain.
On a Mac, point Synopsule at a model running on your own machine. No key, no subscription, and even the transcript text stays put.
On eligible Macs and iPhones with OS 26 or later and Apple Intelligence enabled, summaries, answers, and note context run on-device. Long sources use smaller chunks; summaries are general recaps.
Recording, transcription, speakers, search and exports work without configuring cloud AI.
The honest questions, answered.
How do I get started?
How does Add context work?
What files can I attach?
Do I need a download for search?
Is it one purchase for Mac and iPhone?
Do my recordings sync between my Mac and my iPhone?
Is Synopsule available in my language?
What's different on iPhone?
What does the Apple Watch app do?
Can Synopsule recognize the same person across different recordings?
Where exactly does the audio go?
Do I need an account?
Does it really work offline?
Does the iPhone capture the other side of a call?
How is Synopsule different from cloud notetakers like Otter, Granola, or Fathom?
Can my AI assistant search and control Synopsule?
Can I shape the summary to the kind of conversation?
Does it summarize automatically, or do I have to ask?
Does AI receive my recorded audio?
Does search look inside the transcript, or only at the summary?
speaker: filters. Search also covers notes, labeled flags, and attached text. Results and source quotes help you check where a passage came from. The index is built on your device and can be rebuilt from saved recordings.Will it tell me when a long transcript is done?
Can it know a meeting is starting?
Can I start it with Siri, a Shortcut, or the Action button?
Can it transcribe an audio file I already have?
Is it accessible?
Can I get AI summaries without anything leaving my device?
What do I need to run it?
What formats can I export to?
How much does it cost?
Your next conversation, ready to record.
Open Synopsule, follow the setup, and press record. A speech model is already included. Record and transcribe on Mac or iPhone, or record on Apple Watch and let your iPhone transcribe it. AI is optional.
Get it on the App Store$4.99