One panel for the conversation you're in.

Press H and a floating panel joins your call. It hears both sides through macOS itself - no bot joins the meeting, no account exists, and with the built-in model nothing is uploaded. Connect your calendar and it already knows which meeting this is.

Included with every plan · Free 14-day trial · macOS 26 or later

Assist - answers while you're still talking

Both sides of the conversation are transcribed live. When a question comes your way, a suggested answer appears on its own - or when you press A from any app, so you're never seen clicking into a panel on camera.

Answer one specific line

Hover any line in the transcript and answer exactly that - either speaker, however far back it was said.

One-press rewrites

Make an answer Shorter or Simpler, or ask for More detail or an Example - one click instead of retyping the question.

Or fix it by hand

Edit any result directly - a misheard name, a line that shouldn't be there - instead of regenerating the nine-tenths that were right. The model's own version is kept, so Revert still works in a session you reopen next week, and an exported document says when a human changed it.

Templates for the situation

Save the scenario, who you are and any background - for interviews, sales calls or standups - and reuse it.

Ask about what's on screen

Capture a whole display or just the window in front, always excluding Dictera's own windows, and send it with your next question. One window spends the same pixel budget on a tenth of the area - the difference between a model that can read your stack trace and one that can see there is text there. Needs a model that accepts images.

Meeting Notes - the write-up, without a bot

Same capture, different result. Press Stop and Dictera writes the call up: summary, decisions, action items, open questions. Nothing joins your call and nothing is uploaded when you use the built-in model.

Four levels of detail

Notes, full minutes, a cleaned-up transcript in the language it was spoken, or chapters - a timeline of what was discussed when. Pick which one this meeting is for and both the button and Stop produce it; the other three are one click away.

Shaped to the session

A lecture has no decisions and no owners; a meeting does. Dictera reads each recording for what it is and says which shape it used.

Action items with owners

Name the participants - or let the calendar invite name them - and an item can say who owns it, with "Unassigned" where the meeting never established who, rather than a guess.

Long meetings survive

A transcript longer than the model's window is read in passes, and the result says so when it covers only the most recent stretch.

Everything Meeting Notes does →

Chat - a private model, one shortcut away

The third mode captures no audio at all. It's a typed conversation with whichever model you've pointed it at, including the one already running on your Mac. No browser tab, no account.

Ask without opening anything

Double-tap right ⇧ for a single field in the middle of the screen. Type the question or dictate it, press Return, and the answer is waiting in Chat as a new session. A call the panel is recording carries on untouched.

Nothing is listening

Chat never opens the microphone or the system-audio tap.

Follows the language you type in

Unlike the listening modes, Chat answers in whatever language you write.

Its own engine

Point Chat at the built-in model, a local server or a cloud provider, independently of the other two modes.

Pictures work here too

Capture a display or the front window, or paste an image you already have with ⌘V, and ask about what's in it - same viewer, same rewrites.

Calendar - it already knows which meeting this is

Connect your calendar and a recording arrives knowing what it belongs to, who was invited, and which weekly it's this week's instance of. Everything is read from the accounts already set up in Calendar - iCloud, Google, Outlook, Exchange - so there is no second sign-in. Read-only: Dictera never writes an event, and the reading happens on your Mac.

Off until you turn it on

Calendar access is never requested at first run. Nothing is read until you open Settings ▸ Calendar and switch it on - and you choose which calendars count as sources of meetings.

A countdown before the meeting

About a minute before a meeting starts, a countdown asks whether to record it. Left unanswered it starts recording, because a meeting you didn't record is gone and one you didn't want is a click to delete. Prefer silence to mean "no", or no countdown at all? Both are a setting, globally or for one series.

Nothing ever starts unseen

The countdown is the consent surface. If it can't be put on screen, nothing records - there is no path where Dictera begins listening without showing you first, and the panel is up the whole time it does.

It learns one meeting at a time

Skip the same weekly twice and Dictera stops asking about that one. Stopping a countdown-started recording in the first seconds counts as a skip too - the natural reaction is the signal. Say yes three times and it offers to make that series automatic; it never decides for you, and ignoring a countdown is never counted as a yes.

Attendees fill themselves in

The invitee list becomes the participant names Meeting Notes needs to put owners on action items. Email addresses are never stored at all, and one switch keeps calendar names on local models only.

And a countdown when it ends

A recording that outlives its meeting gets the same treatment in reverse: a minute to say Keep recording, and otherwise it stops - and writes the meeting up on the way out. A recording with nothing scheduled behind it is asked about after an hour or two, which works whether or not you ever connected a calendar.

The day, in the panel

The session list can show the day's agenda - what was recorded, what's still coming and what was missed - with a month grid for reaching any other day. Folded away until you want it.

How it hears the room

Two audio streams, transcribed separately and merged into one two-speaker transcript - you on one side, everyone else on the other.

No Screen Recording permission

The other participants are captured with a Core Audio process tap, so a meeting needs only microphone and audio-recording access - and the call keeps playing normally.

One click mutes you

Both listening modes start with your microphone on, so your own voice is in the call from the first word. Mute whenever you'd rather not be in the transcript; the panel always shows which state you're in.

Speaker echo, cancelled

On loudspeakers the other side would be transcribed twice - once from the call, once as room echo. Dictera gates the mic while they're audible, and stands down when you're on headphones.

Two tabs, one call

Start Meeting Notes while Assist is already listening and the same call is transcribed into both, each in its own session - live suggestions and the write-up, from one conversation.

Pick the device you actually hear on

Capture the participants from your AirPods, or leave it on System Default to follow whatever the output is.

Each mode picks its own model

Something fast and local for live suggestions; something larger for the end-of-meeting write-up. Assist, Meeting Notes and Chat choose independently.

Built in, no setup

Gemma 4 runs on your Mac and covers 140+ languages. Nothing to configure and nothing to send anywhere.

Or your own

Six local servers (Ollama, LM Studio, llama.cpp server, Jan, vLLM, LocalAI) and seven cloud providers (OpenAI, OpenRouter, Groq, Mistral, DeepSeek, Together AI, xAI) are preset. Keys live in the macOS Keychain.

Claude Code CLI, read-only

The Claude Code command-line tool already installed on your Mac - not a Claude model over an API. Point it at a project folder and the assistant can talk about your actual code. There's deliberately no "allow edits": it's driven by what it hears on a call, which is untrusted input.

Connector tools, one at a time

MCP connectors are denied by default. A Jira connector exposes "create ticket" next to "search issues", so nothing is reachable until you allow it individually.

Sessions and series

Everything the assistant produces stays on your Mac - and the meetings you record are mostly the same handful, repeating, so they're organized as sets rather than as a list that only grows.

Searchable session list

Your conversations for the current mode, searchable and titled automatically. Rename one in place in the list, without opening it. Reopening a saved session never stops a recording that's still running.

Filed by the meeting it repeats

Daily, Refinement, Customer interviews - file a session under its series and a year of them stays findable as one set. Filter the list by series, rename and recolor them, and let a template file new ones for you. Nothing is ever labeled automatically - a label that appeared on its own is one you'd have to notice before you could undo it.

What a series still owes

Every open action item across every meeting in a series, in one list. It reads back notes you already have - no model runs, so it's instant and can't invent a commitment nobody made. Where some meetings couldn't be read, it says so rather than quietly showing a short list.

Search marks what it matched

The list, the transcripts, the answers and your templates all paint the characters your search found. It matches the way Finder does, so "resume" finds "résumé" and "OKR" finds "okr" - which matters when you're searching speech in ten languages you didn't type the accents for.

Tidy several at once

Tick a batch of sessions to file them under a series or delete them in one go, instead of one context menu at a time.

Or keep nothing

Transcripts, notes and screenshots live in your Application Support folder and are deleted with the session - or switch assistant history off and nothing is written at all.

The panel itself

It has to sit on top of a video call without becoming part of it. Everything about it is adjustable.

Invisible to screen sharing

Excluded from screen capture by default, and a one-click toggle in the panel rather than a trip to Settings.

18 editable panel keys

Grouped as Actions, Navigation and Display, with Restore Panel Defaults - so nothing clashes with your meeting app.

Answer and Listen from anywhere

A answers, L starts or stops listening - from any app, so you're never seen clicking into a panel. Both are opt-in.

Resize it, or take a column away

Drag any edge, or drag the dividers between the session list, the conversation and the result. Hide the conversation entirely (T) and a long write-up gets the whole window. The size and the spot you left it in come back next launch, and Reset panel layout undoes all of it.

Compact mode

A smaller panel for a crowded screen, adjustable opacity, and answer text you can size up or down. It keeps its own size, so widening the panel to read minutes doesn't follow you onto the next call.

Worth knowing

What it doesn't do.

Two channels means two labels - Dictera doesn't tell individual voices apart, so named attribution comes from the participant list plus whatever the meeting says out loud. Notes are written when you press Stop, not live. Each listening session runs in one language you choose rather than detecting it. No audio is ever stored, so timestamps point at a moment, not a recording. And connecting your calendar never puts a bot in a meeting and never writes to your calendar - it can start a recording on your own Mac, after a countdown you had the chance to answer.

Included with every plan.

The AI Assistant isn't an add-on - it comes with Dictera, and the 14-day trial has no limits on it.