Where to keep your meeting notes

What a meeting notes setup needs

Meeting notes only pay off if you can find them again. A workable setup has three parts: a way to record the meeting and get a transcript, one place where the material lives, and a way to search and ask questions months later. That is the same idea behind a personal knowledge base, and this guide walks through it for meetings.

Step 1: Record the meeting and get a transcript

On the capture side, one option is Pavo, a voice-controlled workspace that records meetings on desktop or mobile, captures mic and system audio, and returns a transcript with speakers, decisions, and action items.

Whatever you record with, aim to finish with a transcript or a short written summary, and a clear list of decisions and actions.

Step 2: Save links and articles in one place

KnowMan saves articles and links, reads and tags them, and keeps them in one library. The browser extension works in Chrome and is listed on the Chrome Web Store.

If an AI agent is connected, the kb_save tool saves a URL to your Inbox or Knowledge Base with an optional note (knowman-mcp). The knowman CLI includes a save command, and it is installed by the same npm package.

Step 3: Search by meaning, not by folder

Weeks later you will not remember the file name. KnowMan's kb_search finds saved items by keyword or by meaning, so a phrase can surface the right item. Titles, URLs, and metadata come back with the results (knowman-mcp).

For how this differs from plain retrieval, see RAG vs a personal knowledge base.

Step 4: Ask questions and get citations

KnowMan answers questions with citations, so each answer points back to the material behind it. Connected agents get the same behavior through the kb_ask tool (knowman-mcp).

The Free plan includes five Ask questions a day (KnowMan). The pattern behind this is covered in what an AI knowledge base does.

Keep the decisions, not just the transcript

Meetings produce decisions, and decisions are what you reuse later. KnowMan also keeps memories: claims, decisions, and playbooks (KnowMan). The kb_memories tool serves them to connected agents (knowman-mcp), which is close to persistent memory for AI agents.

Step 5: Connect your AI agent over MCP

KnowMan ships an MCP server, so an agent can search, read, ask, and save in the same library. It runs locally as a stdio server started with npx -y @knowman/mcp on Node 22 or newer (npm), and as a hosted remote server at https://api.knowman.io/mcp with OAuth sign-in or an API key (knowman-mcp). Per the knowman-mcp project, the hosted server is verified with the official MCP SDK client.

Agents get six tools: kb_search, kb_ask, kb_get_item, kb_save, kb_recent, and kb_memories (knowman-mcp). Setup walkthroughs live in the MCP server guide, and connecting any AI to your notes covers the wider field.

Where KnowMan fits

Pavo covers the meeting itself: the recording, the transcript, and the decisions and actions you keep from it (Pavo). KnowMan covers the library side: it saves articles and links, reads and tags them, answers questions with citations, and serves that library to AI agents over MCP (KnowMan). For agents in particular, start with a knowledge base for AI agents.

Pavo describes the recording side in turn meeting transcripts into a searchable knowledge base.