QMD

Overview

QMD is a search and retrieval concept referenced as part of AI-assisted workflows that rely on a large personal or operational knowledge base. In the newsletter mentions, it appears in two practical contexts: as a memory index used with OpenClaw’s Felix agent, and as the search layer Ryan Wiggs used to ingest roughly 5 million words of prior PM work into Claude Code. In both cases, QMD functions less like a standalone end-user app and more like an enabling retrieval system for finding relevant context across a large corpus.

For AI Product Managers, QMD matters because it points to a core design pattern in modern AI systems: durable memory plus targeted retrieval. Rather than asking an LLM to operate from stateless prompts alone, teams can use a search layer like QMD to surface prior documents, decisions, research, and artifacts at the moment they are needed. That can improve continuity, reduce repeated work, and make AI tools more useful for strategy, execution, and knowledge reuse.

Key Developments

• 2026-02-23: QMD is referenced as a memory index in Nat Eliason’s OpenClaw workflow, where Felix uses a QMD-based memory layer alongside APIs, automation, and a cron-driven heartbeat to help autonomously launch and operate a business workflow.
• 2026-04-23: QMD is described as the search mechanism Ryan Wiggs used to ingest five years of PM work, about 5 million words, into Claude Code to create a “second brain” that reportedly doubled productivity.

Relevance to AI PMs

• Build AI products with retrieval, not just prompting. QMD highlights the importance of pairing models with a search layer over internal docs, research, specs, user feedback, and decision logs so outputs are grounded in real product context.
• Turn PM artifacts into reusable institutional memory. Product reviews, PRDs, strategy docs, postmortems, and customer notes often become dead documents. A QMD-style retrieval system can make them queryable inside AI workflows, increasing leverage for PMs and their teams.
• Prototype “second brain” workflows for execution speed. AI PMs can test systems where coding agents, writing assistants, or research copilots pull from a curated knowledge base to draft plans, answer historical questions, and maintain continuity across long-running initiatives.

Related

• nat-eliason: Referenced using a QMD-based memory index in an OpenClaw business-building workflow.
• openclaw: One of the systems where QMD appears as part of the memory and retrieval stack powering an autonomous agent.
• memory-system: QMD is closely related to the broader idea of persistent AI memory, especially retrieval over accumulated documents and prior work.
• ryan-wiggs: Described using QMD search to ingest years of PM output into Claude Code as a productivity-enhancing “second brain.”
• claude-code: QMD is mentioned as the retrieval layer feeding historical PM knowledge into Claude Code workflows.