LoopX Upgrades to 1.0: Manage Long-Running Agents like Codex and Claude Code from a Single Page

动察 Beating AI Flash: Open-source project LoopX has released version 1.0. It sits atop agents like Codex, Claude Code, and Cursor, consolidating long-running tasks previously scattered across different sessions into a single unified workspace.

Users can view all tasks connected to LoopX on one page. Which ones are executing, which are awaiting confirmation, which are under continuous monitoring, and which are scheduled—all are displayed in a centralized view. After entering an individual project, users can further inspect the to-do lists, completion records, files, and operational status of different agents.

This workspace can be opened directly in a browser and also offers macOS and Windows desktop versions. Both entry points share the same local service and task state. It does not automatically read all agents on your computer; instead, it displays only the agents and tasks that have been registered and connected to LoopX.

Multiple agents can also collaborate to achieve a single goal. They can each claim tasks, work in parallel, and hand off to one another. For example, after one agent finishes, it leaves follow-up tasks for the next agent to pick up. This collaboration is driven by shared tasks and state, rather than having multiple agents freely converse in a chat room.