ИИ-агенты
zeron
Native control plane for Claude Code, Codex, Cursor, Devin, and other coding agents.
★ 3,0 тыс. звёздMIT
Смотреть руководство →loopx — проект с открытым исходным кодом из категории «ИИ-агенты» от loopx-project (loopx-project/loopx на GitHub). У него 6 208 звёзд на GitHub и 596 форков, он распространяется по лицензии Apache-2.0 и последний раз обновлялся на GitHub 8 октября 2026.
A control plane with a durable state kernel that keeps long-horizon agents moving across sessions.
Define the multi-session outcome you want an agent to keep advancing, plus the boundaries it must not cross.
Install the Python package and its workflow skills.
Run LoopX in your agent host session; it connects the project, plans the work as Todos and sets up the heartbeat.
Start with /loopx or $loopx in your host, then use loopx status to see the goal, pending decisions and the next Todo.
Goals and progress persist across sessions, so long-horizon work survives restarts and context resets.
A heartbeat keeps connected work moving and improving over time with bounded continuation, instead of stopping after one turn.
Start from Claude Code (/loopx), Codex ($loopx) or the DeepSeek Harness plugin; LoopX reuses your existing project state.
Work is planned as Todos; loopx status shows the goal, any decision waiting for you and the next step, and loopx doctor diagnoses installs.
LoopX can generate bounded turn decisions from validated receipts, quota state and a provider-neutral budget.
Safe presets cover recurring jobs such as daily triage, changelog drafts and PR watching, with readiness scoring for repeated runs.
Clone LoopX if you want the source; most users install the package in the next step.
git clone https://github.com/loopx-project/loopx.git
cd loopxInstall the package from PyPI and install the workflow skills.
python3 -m pip install --upgrade loopx && loopx workflow-skills --installRestart your agent host, open a session on your project, and hand LoopX a complex task.
# In Claude Code: /loopx <complex task>
# In Codex: $loopx <complex task>
loopx statusLoopX is maintained by the loopx-project organization, with documentation on its project site.
Yes — it is open source under the Apache-2.0 license. You still pay for whatever agent host and models you run it with.
LoopX itself is a control plane, not a model; it works through your existing agent host (Claude Code, Codex, DeepSeek Harness), which supplies the model access.
The hosts execute turns; LoopX adds the durable goal state, heartbeat continuation, todos and governance that keep long-horizon work coherent across sessions.
pip install loopx, install the workflow skills, restart your agent host, and run /loopx with a complex task in your project session.
Native control plane for Claude Code, Codex, Cursor, Devin, and other coding agents.
A benchmark testing whether agents can solve long-horizon coding tasks.
Self-hosted MCP gateway and control plane — connect, control and operate all your MCP servers from one place
Uber's ADR: discover, observe, benchmark, and detect threats across enterprise AI agents, deployed at Uber.
Agentic framework for long-horizon autonomous scientific discovery.
Durable, composable AI agents — flows that crash and resume, retry safely, and explain every step.
Другие проекты из категории «ИИ-агенты» с наибольшим числом звёзд в каталоге LibreAI — реалистичный список альтернатив, если loopx вам не подходит.
A personal AI that really does things — on any OS, with your data staying yours.
An agentic skills framework & software development methodology that works.
Agent harness optimization system with skills, memory and security for Claude Code, Codex and more
The agent that grows with you — a Python framework for adaptive AI agents.
DeepSeek's plugin-based agent harness where everything is a plugin.
Соседние проекты из категории «ИИ-агенты» в каталоге, ближайшие по числу звёзд к loopx — продолжайте изучать полку по обе стороны от него.
A local-first AI agent workspace for coding, writing, design, research and automation.
Open-source LLM observability: monitor, evaluate, and experiment.
A Large Action Model framework for building AI web agents that browse and act for you.
An Agent OS with interview-gated, budgeted evolution — one MCP server across 14 agent runtimes.
Policy, identity, and sandboxing guardrails for AI agents.
Open-source AI coworkers, each with a computer of their own.
Бесплатно и открыто — забирайте исходный код, ставьте звезду репозиторию и начинайте создавать.