wtf is loopx?

LoopX is a tool that helps manage AI coding assistants, like Claude Code or Codex, when they are working on tasks that stretch across many hours or days rather than a single short session. Normal AI agents tend to lose track of goals, todo lists, and context once a session ends or a conversation gets too long. LoopX keeps that information in a separate, durable layer outside the AI model itself, so work can be reviewed, paused, and picked back up later without starting over. It tracks things like the overall objective, checklists of gates that must pass before moving forward, a list of concrete todo items, evidence of what has actually been done, and a "quota" system that governs how much the AI agent is allowed to keep working before checking back in with a human. When a decision genuinely requires a person's judgment, LoopX is designed to pause and ask a specific question rather than guessing. When a safe next step is clear, it lets the AI agent run one bounded chunk of work at a time. LoopX describes itself as a control layer that sits alongside an existing AI coding tool rather than replacing it, and it works with multiple different agents including Claude Code, Codex, and Cursor, instead of being tied to one company's product. It is aimed at situations like long multi-day engineering projects, ongoing research work, ticket and pull request tracking that needs to preserve context, or recurring monitoring tasks. The project includes documented examples of open source contributions and research experiments that ran for over 200 hours using this system to keep track of state. Installation and usage details point to a public documentation site and a getting started guide. The project is released under the MIT license, which allows free use, modification, and commercial use as long as the original copyright notice is kept.