---
title: Blog
summary: Meet Holon, explore its design, and see how long-lived agents work in practice.
order: 4
---

<div class="blog-page">
<header class="blog-intro">
<p class="blog-eyebrow">THE HOLON BLOG</p>

# Work that keeps going.

<p class="blog-intro__lede">The ideas behind Holon, the design that supports them, and the people putting agents to work.</p>
</header>

<a class="blog-feature" href="/blog/what-is-holon">
<img src="/assets/holon-agent-workspace-cover.webp" width="1672" height="941" alt="" decoding="async" fetchpriority="high">
<div class="blog-feature__copy">
<span class="blog-category">START HERE · PRODUCT</span>
<h2>What is Holon? Let agents keep working in your environment.</h2>
<p>Give recurring work to agents with ongoing roles. A tour of the local workbench, from continuous review to remote development and team use.</p>
<span class="blog-read">Meet Holon <span aria-hidden="true">→</span></span>
</div>
</a>

<section class="blog-stories" aria-labelledby="blog-stories-title">
<div class="blog-section-heading">
<h2 id="blog-stories-title">Design &amp; practice</h2>
<p>How it works. How to use it.</p>
</div>
<a class="blog-story" href="/blog/why-work-items">
<img src="/assets/workitem-abstract-flow-cover.webp" width="1536" height="768" alt="" loading="lazy" decoding="async">
<div><span class="blog-category">DESIGN</span>
<h3>WorkItems: separating work state from conversations</h3>
<p>An API migration and two PRs show why goals, progress, and waiting conditions need a home of their own.</p>
<span class="blog-read">Explore the design <span aria-hidden="true">→</span></span></div>
</a>
<a class="blog-story" href="/blog/one-pr-one-work-item">
<img src="/assets/continuous-pr-reviewer-cover.webp" width="1672" height="941" alt="" loading="lazy" decoding="async">
<div><span class="blog-category">HANDS-ON GUIDE</span>
<h3>Build a reviewer that follows PRs through</h3>
<p>Set up the code-reviewer template, agree on permissions, and subscribe to PRs. Follow revisions and CI through to the merge decision.</p>
<span class="blog-read">Set up continuous review <span aria-hidden="true">→</span></span></div>
</a>
<a class="blog-story" href="/blog/agents-in-a-small-team">
<img src="/assets/team-shared-agents-cover.webp" width="1536" height="1024" alt="" loading="lazy" decoding="async">
<div><span class="blog-category">TEAM STORY</span>
<h3>From personal AI tools to shared team agents</h3>
<p>A small team's 77 days of recorded usage, device investigations, and hands-on testing show where shared agents fit—and where people still step in.</p>
<span class="blog-read">Read the team story <span aria-hidden="true">→</span></span></div>
</a>
</section>

<footer class="blog-next">
<div><h2>Try it with one responsibility.</h2><p>Install Holon, connect a provider, and give your first agent a bounded job.</p></div>
<a href="/getting-started/">Get started <span aria-hidden="true">→</span></a>
</footer>
</div>

<!-- INDEX:START -->

- [What is Holon? Multiple agents working in your environment](./what-is-holon.md)
  Assign recurring work to established roles, let agents follow up in the background, and reconnect when needed. An ongoing review example introduces Holon's local workbench for remote development and team collaboration.
  <!-- mdorigin:index kind=article -->

- [Holon's WorkItem architecture: separating work state from conversation](./why-work-items.md)
  Follow an interface migration from launch to retirement to see how WorkItem differs from Goal and Loop: it keeps objectives, progress, external dependencies, and delivery together across release cycles.
  <!-- mdorigin:index kind=article -->

- [Beyond a one-time review: build a reviewer that follows PRs with Holon](./one-pr-one-work-item.md)
  Create a reviewer with built-in working rules and Skills, set its responsibilities and merge permissions, and have it discover PRs and follow fixes and CI updates.
  <!-- mdorigin:index kind=article -->

- [From personal AI tools to team collaboration: an agent-native case study](./agents-in-a-small-team.md)
  How a small team moved from individual AI tools to shared agents, with 77 days of retained usage records, device investigations, and a hands-on test of silent piano keys.
  <!-- mdorigin:index kind=article -->

<!-- INDEX:END -->