Open-source local hub that turns command-line coding agents into a background loop: it takes work from webhooks, issue polling, cron rules or an HTTP API, dispatches it to Claude Code, Codex, ACP agents or any CLI, and only sends you a decision card when your call is needed.
Follow Agent Autoloop to get an email when its listing changes, such as new pricing or features. One digest a day at most, and you can unsubscribe anytime.
- 01
Zero-dependency, event-sourced SQLite kernel (about 1,650 lines)
- 02
Agents: Claude Code, Codex, ACP, DeepSeek Harness or any prompt-in/reply-out CLI
- 03
Events from GitHub/GitLab webhooks, issue polling, cron rules and an HTTP API
- 04
Decision cards via desktop notifications, Feishu, DingTalk, WeCom, Slack or Discord
- 05
Zero-cost wait: no agent process or tokens while waiting for your answer, then the session resumes
- 06
Hop circuit breaker and `agent-autoloop why` run histories
- 07
`deliver` label pipeline: review, wait for CI, merge, wait for deploy, verify
Open Source
Free and open-source (MIT). Runs locally on your machine with the CLI agents you already have; model usage is billed to your own provider account.
Best for
Developers on macOS or Linux who want their existing Claude Code or Codex setup to pick up issues, alerts and scheduled chores in the background with human approval only when needed.
Not ideal for
Windows users (native support pending; WSL2 only), teams wanting a hosted dashboard, or users who need English CLI output and docs today.
A thoughtfully engineered, fast-growing way to run CLI agents unattended with human-in-the-loop cards; very new and Chinese-first, so expect rough edges for English-speaking users.
v1.3.0 through v1.4.1 shipped on Nov 5, 2026, two days after the Nov 3 launch, with kernel and plugin cleanups; native Windows support is promised in issue #17.
As of
Is Agent Autoloop free?+
Yes - Agent Autoloop is Open Source. Free and open-source (MIT). Runs locally on your machine with the CLI agents you already have; model usage is billed to your own provider account.
Is Agent Autoloop open source?+
Yes - Agent Autoloop is open source. Free and open-source (MIT). Runs locally on your machine with the CLI agents you already have; model usage is billed to your own provider account.
Who is Agent Autoloop best for?+
Developers on macOS or Linux who want their existing Claude Code or Codex setup to pick up issues, alerts and scheduled chores in the background with human approval only when needed.
Who is Agent Autoloop not ideal for?+
Windows users (native support pending; WSL2 only), teams wanting a hosted dashboard, or users who need English CLI output and docs today.
What are the best Agent Autoloop alternatives?+
The closest Agent Autoloop alternatives on ai.dosa.dev are Jules by Google, OpenHands, SWE-agent - all listed under Autonomous & Async Agents.
Who makes Agent Autoloop?+
Agent Autoloop is developed by Khazix. It is listed in the Autonomous & Async Agents category on ai.dosa.dev.
Favorite this tool to revisit it later, or Zap it to contribute to the public vote count.