2026. Active.
Persistent memory for AI coding agents. Findings, tasks, and patterns stay as markdown in a git repo you own, and you can browse the same store from a terminal shell, a 3D web graph, VS Code, or an iPhone.
Agents lose context between sessions. Phren keeps it in a git repo you own.
AI coding agents lose everything between sessions, so the same context has to be rebuilt over and over.
I built it as a TypeScript monorepo: an MCP server that exposes ten tools by default (an admin tool reaches the rest by name, so the schema stays small; the full profile exposes all 77), a CLI, and hooks that capture findings and tasks into plain markdown. No database, no vendor lock-in. Findings have a lifecycle (supersede, retract, contradict) and trust that decays over time, and a fragment graph links concepts across projects. The store has seven front ends on top of the same files: a full-screen terminal shell, a braille-canvas terminal graph that lights up as agents read and write, a 3D web viewer, a VS Code extension, a SwiftUI iOS app with widgets and Siri intents, and Herdr and Omarchy plugins.
1. In the terminal, pick a fragment and see every project that links to it.
2. Tasks for each project, kept as plain markdown: active, queue, done.
3. The same store in the browser, as a graph you can move through.
It works across Claude, Copilot, Cursor, and Codex, installs as a Claude Code plugin, syncs shared stores over git, and reloads cleanly on a new machine with a single init command.
Findings, tasks, and context persist as markdown across sessions, projects, and machines, and reload automatically.
Made with TypeScript, MCP, SwiftUI and Node.js.