Be in the conversation.
Remember what matters.

No. Kesami records on your computer: your microphone for you, the meeting’s audio for everyone else. Nobody on the call gets an invite, a join request or a bot in the participant list.

Yes, treat it as your call to make before you press record. Because no bot joins, Kesami does not announce itself to the other people on the call, so tell them you are taking notes and get whatever permission your meeting, client or local rules require. Kesami does not give legal advice, and the rules differ by place and by who is on the call.

Only the services that do the work you asked for. While you record, audio streams through Kesami’s relay to Sarvam for live transcription, and the relay keeps no copy of the audio or the meeting. When you generate a summary or ask a question, the transcript or the relevant excerpts go to Google Gemini and, if Gemini is at capacity, OpenAI. Those providers’ own terms govern how they process and retain what they receive, and Kesami does not promise they never keep it. Your recordings, transcripts and notes stay in a folder on your computer.

Anything that plays sound on your computer — Google Meet, Zoom, Microsoft Teams, Slack huddles, even a recorded video. Kesami numbers the voices it hears; rename a speaker once and every line follows, with the calendar invite’s attendees suggested as names.

Macs with Apple silicon (M1 or later) on macOS 13 Ventura or newer, and 64-bit PCs on Windows 10 or 11 (beta). There is no Intel Mac or Windows on Arm build yet. Check: Apple menu › About This Mac shows Chip: Apple M…

Windows support is in beta, so expect rough edges. The main difference is how Kesami hears the other side of the call. On a Mac it captures the meeting’s sound directly. On Windows it does that when you record with Screen + sound; for audio-only meetings, pick a loopback input such as Stereo Mix or VB-Audio Cable under Settings › Audio › Meeting audio. Your own microphone works the same everywhere. Report what broke

On a Mac, yes. Turn on Voice typing from Kesami’s Home screen and allow the Accessibility, Microphone and Speech Recognition permissions it asks for. In macOS Keyboard settings, set ‘Press 🌐 key to’ to ‘Do Nothing’ so the key only talks to Kesami. Then hold Fn, speak and let go: the text is pasted where your cursor is, and your clipboard is put back afterwards, except PDF, rich text with attachments, audio and file-promise contents, and a very large copy may come back only in part. Double-tap Fn, or press Fn and Space, to dictate hands-free, and press Fn again to finish; Esc cancels. Hold Fn and Control to give an instruction instead: with text selected, say what to change, for example “make this more formal” or “translate this into Hindi”, and the result replaces the selection; with nothing selected, describe what to write. Speech recognition runs on your Mac, and on macOS 26 or later Hindi can be recognised after a one-time model download from Apple. On macOS 26 or later with Apple Intelligence turned on, a clean-up pass on your Mac also removes filler words and self-corrections and fixes punctuation, in the languages Apple Intelligence supports. Voice commands and Ask AI send only the selected text and your instruction to the same AI that writes your summaries, use the Free plan’s 20 monthly AI uses (the same allowance as chat replies), and nothing is saved as a meeting. In apps that don’t share their selection, Kesami copies it with ⌘C; editors that copy the whole line when nothing is selected send that line. If a field doesn’t take the paste, the text stays on your clipboard. Password fields are refused. Voice typing is part of the Mac app; the Windows (beta) build doesn’t include it.

Mac: Signed with an Apple Developer ID and notarized by Apple. It opens like any other app. Windows (beta): the installer isn’t code-signed yet. When SmartScreen says “Windows protected your PC”, click More info, then Run anyway. Kesami checks for new versions while it runs: it downloads the update in the background and installs it when you restart, on both Mac and Windows.

On your computer, in Documents › Kesami Meetings — one folder per meeting, with the transcript and summary as Markdown and the details as JSON. While you record, audio streams through Kesami’s relay for live transcription. When you choose Generate, the transcript goes to AI to write the summary. Asking a question sends the most relevant retrieved excerpts for an answer.

Keep recordings in your own Google Drive or Dropbox. Uploads resume after interruptions, playback streams from your account, and verified uploads can free local disk space. Keep selected recordings offline. Optional and off by default. Storage capacity depends on your cloud account’s quota. Cloud playback requires an internet connection. Set this up in Settings → Storage; your meeting notes stay on your computer.

English, Hindi, Bengali, Gujarati, Kannada, Malayalam, Marathi, Punjabi, Tamil, Telugu and Urdu — or let Kesami detect the language for you. Transcription runs on Sarvam’s models, speech models built in India for Indian languages and for sentences that switch between Hindi and English.

Kesami transcribes with Sarvam, whose models are built for Indian languages and mixed Hindi-English speech; Granola lists Hindi among 31 languages through its transcription provider. Kesami keeps every meeting as files on your computer; Granola stores notes and transcripts on its servers. Kesami Pro is ₹499 or $10 a month; Granola Business is $14 a user a month. Granola has phone and Apple Watch apps, team workspaces, and SSO on Enterprise; Kesami does not yet. Read the full comparison

Yes, a Google sign-in. Your free hours and your Pro plan are attached to it, and it lets Kesami transcribe and summarise through its own relay instead of asking you for API keys.