"Nota Mesyuarat Saya Bertaburan Merentas Lima Aplikasi"

Think about the last recurring project you worked on. Weekly syncs, stakeholder reviews, design critiques — maybe thirty meetings over two months. Where do those notes live? Some in Notion. Some in your personal notes from the first three weeks when you were being organized. Some in a Google Doc that someone else started and stopped maintaining. Most of it lives in people's memories, which erode in proportion to how busy things get. When a new hire joins in month three, "getting up to speed" means interrupting five people for half a day each.

The pattern we kept running into is this: the raw meeting content is gold — decisions, rationale, constraints, people's positions — but nobody has time to synthesize it into anything durable. The individual meeting note is fine for a week. The project-level source of truth almost never exists. Andrej Karpathy wrote a widely-shared note about this exact gap: an AI-native knowledge base that updates itself from every conversation, organized as a living wiki of decisions, people, glossary, and how the project evolved — rather than a dumping ground for raw transcripts. That note shaped how we built the centralized-notes feature.

Every session you tag with a topic feeds that topic's wiki. The AI extracts decisions with the reasoning behind them, identifies the people involved and their roles, builds a glossary of project-specific terms, and tracks insights with evidence. Before your next meeting, open the topic and read the evolution card — a two- or three-sentence briefing on where things stand and what changed since last time. Need to answer "what did we decide about the third-party vendor?" Ask the Q&A bar and get an answer sourced directly from the transcripts of your actual meetings. The knowledge base grows quietly in the background, session after session, and is still useful a year later when you need to remember why a decision was made the way it was.