Repolode
Submit
loopx-project

loopx-project/loopx

A local-first control plane that keeps goals, decisions, ownership and evidence durable for long-running agent work across sessions and runtimes.

View on GitHubWebsite
Stars
6,560
Forks
615
Open issues
177
Last commit
1 hour ago

What is it?

LoopX is a control layer for long-running work done by AI agents. An agent can finish a single task in one session, but multi-day engineering, research or operations goals suffer when objectives change, owner decisions arrive and evidence goes stale. LoopX stores the Goal, accepted requirements, decisions, todo ownership, gates and evidence locally, so an agent or a person can reopen a project and see what changed and what to do next.

LoopX does not supply the model or tools itself. It works alongside runtimes such as Codex, Claude Code, DeepSeek Harness, KunlunCode, OpenCode and others through host-specific install commands, and it can be driven from a custom runner using a small CLI loop (quota checks, claims, updates, state refresh). Protected actions such as publishing or production writes remain under human authorization. The project also ships a browser dashboard, desktop previews, a research workbench for tracking explorations and findings, and optional capabilities for issue fixing, periodic reports and benchmark evidence.

Key features

  • 01Keeps goals, decisions, todo ownership and evidence in durable local state across sessions and restarts.
  • 02Integrates with Codex App and CLI, Claude Code, DeepSeek Harness, OpenCode and other agent hosts.
  • 03Provides a CLI loop (quota check, claim, update, refresh, spend-slot) for custom runners.
  • 04Requires human authorization for protected actions such as publishing or production writes.
  • 05Includes a browser dashboard and a research workbench for tracking explorations and findings.

Use cases

  • Keeping a coding agent working on a multi-day refactor or performance task across sessions.
  • Running long research or benchmark experiments with recorded decisions and evidence.
  • Coordinating several agents that hand off work and review each other's output.
  • Attaching a durable goal and approval gates to a recurring agent loop in a custom runner.

Quick start

  1. 1

    Install LoopX and its workflow skills with pip

    python3 -m pip install --upgrade loopx && loopx workflow-skills --install
  2. 2

    Restart your Agent host, open your project, and give it a concrete task (for example, Codex: $loopx <task>)

  3. 3

    Open the Personal Agent Workspace in your browser

    loopx dashboard

Things to know

  • Requires Python 3.11+ and Node.js 22.22.3+ (24 LTS recommended).
  • Desktop App builds are previews with platform limits: the macOS App is ad-hoc signed and not notarized, and the Windows installer may trigger a SmartScreen prompt.

In collections

Claude Code tools, skills & pluginsOpen-source AI agent frameworks & appsAI coding assistants & agentsPrivacy & local-first tools

Topics

#ai-agents#agent-control-plane#long-horizon-agents#multi-agent#local-first#claude-code#codex#agent-harness#agentic-software#auto-research

FAQ

What is loopx?
A local-first control plane that keeps goals, decisions, ownership and evidence durable for long-running agent work across sessions and runtimes. LoopX is a control layer for long-running work done by AI agents. An agent can finish a single task in one session, but multi-day engineering, research or operations goals suffer when objectives change, owner decisions arrive and evidence goes stale. LoopX stores the Goal, accepted requirements, decisions, todo ownership, gates and evidence locally, so an agent or a person can reopen a project and see what changed and what to do next.
Is loopx free to use? What license does it have?
loopx is open source under the Apache-2.0 license. Check the license terms before using it in commercial products.
Is loopx still maintained?
Actively maintained. The last commit to loopx was on Nov 7, 2026.
What can I use loopx for?
Keeping a coding agent working on a multi-day refactor or performance task across sessions. Running long research or benchmark experiments with recorded decisions and evidence. Coordinating several agents that hand off work and review each other's output. Attaching a durable goal and approval gates to a recurring agent loop in a custom runner.
How do I install loopx?
Install LoopX and its workflow skills with pip: python3 -m pip install --upgrade loopx && loopx workflow-skills --install Restart your Agent host, open your project, and give it a concrete task (for example, Codex: $loopx <task>) Open the Personal Agent Workspace in your browser: loopx dashboard
Where can I get loopx?
The source code and installation instructions are on GitHub: https://github.com/loopx-project/loopx

Summary generated by AI from the project's README. GitHub stats refresh daily.

Related projects