Topic
7 pieces
Reference Architecture · Agents
Agent memory over MCP puts an assistant's long-term memory in its own small server, so that any app that speaks the Model Context Protocol can save facts and look them up again. Explained with three diagrams and one worked example.
Demo · MCP
A local MCP server that lets Claude Desktop, Cursor or your own agent remember facts and recall them by meaning later. Each tool maps to a CoALA memory action.
Demo · Voice AI
An open-source reference pipeline for speech-to-text followed by LLM inference, with every model claim checked against the transcript before it is shown. The worked example is call-centre quality review; each stage maps to the CoALA framework.
Reference Architecture · Agents
A reference architecture for thinking about any LLM agent: three kinds of long-term memory, a working memory where the model reasons, an action space, and a decision cycle that ties them together. Explained with two diagrams and one worked example.
Demo · RAG
A minimal web app that shows semantic Retrieval-Augmented Generation end to end. Pick a knowledge base (HR policies, a lease, a router manual), ask a question, and see the passages, similarity scores and database rows behind every answer. Each stage maps to the CoALA framework.
Free Skill
The sibling of demo-doc for ideas rather than codebases: "what is an agent harness?", "what is AI Ops?". Seven tutorial-style sections, two or three hand-drawn diagrams, links to the demos that implement it, and credit to every original source.
Free Skill
A Claude skill that writes a demo's design document the same way every time: ten numbered sections, a CoALA pipeline strip, hand-drawn figures and fork-and-contribute links, as DESIGN.md, a styled HTML page and a website post.