exocortex

Static wiki engine written in Uxntal

Context as a structured knowledge layer in plain text.

Various and sundry additional pieces of software I've written to incorporate into my exocortex.

Exocortex template: fork & deploy your AI-powered personal knowledge system with Claude Code

Give Claude Code a persistent memory — it remembers you, your projects, and your decisions as plain Markdown files you own. Fully local. Just say `start`.

One prompt that grows into your own AI second brain — a lightweight seed an agent reads to build a personal knowledge base shaped to you.

Context vault and deployment

Personal website, come take a look!

The Exocortex of 2130 -- version 1.0

Personal AI exocortex — multi-agent LLM system on GCP: long-term memory, realtime voice companion, semantic search over Google Drive and OneDrive, agent skills, MCP connectors for Claude. Hexagonal architecture, provider-agnostic (Claude/Gemini/OpenAI/Grok).

Creative writing, essays, and philosophical explorations from the Exocortex project

MCP server and local sync package for Cosmos. Connects agents, iMessage, browser history, calendar, Claude Desktop, and shell history to a shared personal graph.

Local-first cognitive workspace implementing Agent2UI (A2UI), Agent2Agent (A2A) and MCP protocols for recursive human-AI collaboration. Next.js based, offline-capable, and fully inferred locally.

An AI chief of staff that reads your stuff, remembers what matters, and runs your life in the background.

My (public) notes implemented with VSCode's "Foam" extension, as well as for use with any Markdown tools

The Exocortex interface documents — OpenAPI and AsyncAPI generated from the server's own routes and channels, plus federation-v1.md, which is written by hand and before the code. CC0, so a second implementation of the protocol owes nothing to this one.

A Model Context Protocol server that gives an agent read and write access to a local Exocortex node — PARA buckets, blocks, links and tags, plus the workspace's own conventions so the agent files things where you would. Stateless, local only; there is no hub endpoint.

Shared GitHub Actions for the Exocortex repositories — go-check, node-check, go-release and codeql, called by every product repo in about fifteen lines. Six near-identical workflow files would be six places to fix the same bug, and they drift on the first fix.

Rust crate: memory-plimpsest

The Exocortex front end — one renderer shared by three shells, and a single bridge module that knows anything native. Electron on the desktop, a plain SPA on the web, Capacitor over an embedded Go node on Android. Talks to a node or a hub through generated API types.