Plain-English definitions of the words behind AI meeting notes — bot-free recording, citation-backed summaries, diarization, RAG chat, and more.
AI meeting notes are summaries, transcripts, and action items generated automatically from a recorded meeting by an AI system. Instead of typing notes during a call, you record it and the AI produces a speaker-labeled transcript and a structured summary afterward. The best tools link each summary point back to the transcript moment that supports it.
An AI notepad is a note-taking app with a built-in AI layer that captures meetings, transcribes them, and writes summaries you can edit alongside your own notes. Unlike a plain recorder, it is a workspace: you type, the AI fills in the transcript and summary, and the two live together in one document.
Bot-free meeting recording captures a call without sending a separate bot participant into it. Instead of a "Notetaker has joined" bot, the app records your microphone and the meeting's system audio directly from your own device. Nothing announces itself to the other side, so the conversation stays unannounced and private to you.
A meeting bot, or notetaker bot, is an automated participant an AI notetaker sends into a video call to record and transcribe it. It appears in the participant list — often as "Otter.ai" or "Fireflies Notetaker" — which signals to everyone that software is recording the meeting.
A citation-backed summary is an AI meeting summary in which every point links to the exact transcript moment that justifies it. Instead of trusting that the AI summarized correctly, you click a claim and check the source. It is the antidote to AI summaries that paraphrase inaccurately or hallucinate details.
Speaker-attributed transcription is a transcript that labels who said each line, not just what was said. Attribution can be by named speaker (via diarization) or by audio source — for example, separating your microphone from the meeting's system audio to distinguish you from everyone else on the call.
Speaker diarization is the process of partitioning a recording by speaker — answering "who spoke when." It groups the audio into distinct voices and labels each transcript line accordingly. Diarization is harder on noisy or overlapping speech, which is why some tools instead separate speakers by audio channel.
System audio capture records the sound your computer plays — the other participants' voices in a video call — rather than only your microphone. Combined with mic capture, it lets a recorder reconstruct the full conversation locally on your device, without a bot joining the call to collect the remote audio.
RAG (retrieval-augmented) chat lets you ask questions across your meeting notes and get answers grounded in the actual transcripts. The system retrieves the relevant passages from your notes and folders, then generates an answer that cites them — so "what did this account say about pricing?" returns a sourced answer, not a guess.
Real-time transcription converts speech to text live, as the meeting happens, rather than after it ends. The words stream onto the screen within a second or two of being spoken, so you can read the conversation as it unfolds and catch anything you missed without waiting for processing.
Local recording, or on-device capture, means the audio is recorded on your own computer rather than by a server or a bot in the call. Note that local capture does not always mean local processing — many tools record on-device but still send the audio to the cloud for transcription and AI summaries.
An AI meeting assistant is software that records or joins your meetings and automates the busywork around them — transcription, summaries, action items, and follow-ups. Some assistants attend as a bot; others capture bot-free from your device. The category overlaps heavily with "AI notepad" and "AI meeting notes."
A meeting Lens is a reusable AI template that shapes how a meeting is summarized — a sales-discovery Lens, a design-review Lens, or a standup Lens. Applying a Lens to a transcript produces a consistent, structured output every time, so different meeting types get the format that actually fits them.
Action items are the concrete to-dos an AI extracts from a meeting — who agreed to do what, by when. A good AI notetaker pulls them from the conversation automatically and can push them to tools like Slack, Linear, or Notion so the follow-ups do not get lost after the call.