构建始终在线的智能体
Over the past few months, we’ve seen a resurgence of always-on agents.
在过去几个月里，我们看到了始终在线的智能体的复兴。
Always-on agents are stateful and proactive, meaning they manage context across sessions and take actions without direct supervision. Underneath, they run on the same core loop as other agents. The difference is the surrounding architecture, which is built from primitives that control when the agents wake up, what carries over between runs, and how their work reaches the user.
常驻智能体是有状态且主动的，这意味着它们跨会话管理上下文，并在没有直接监督的情况下采取行动。在底层，它们与其他智能体运行在相同的 核心循环 上。区别在于周围架构，它由原语构建，这些原语控制智能体何时唤醒、运行之间保留什么，以及它们的工作如何到达用户。
The resurgence is largely due to recent model advancements, which have made always-on agents more practical to run. First, the latest models have better judgment: they’re more reliable at understanding user intent and knowing when to check in. Second, intelligence is becoming cheaper, with long-horizon tasks benefiting the most. For example, on one long-horizon coding benchmark, Opus 5.5 cost 78% less per solved task than Opus 5, compared with 40% less on typical workloads.
这种重新兴起很大程度上归因于近期的模型进展，这些进展使始终在线的智能体更易于实际运行。首先，最新模型具有更好的判断力：它们在理解用户意图以及知道何时进行检查/汇报方面更可靠。其次，智能成本正在降低，长周期任务受益最大。例如，在 一个长周期编码基准测试 中，Opus 5.5 每个已解决任务的成本比 Opus 5 低 78%，而典型工作负载下则低 40%。
Customers have started to ask us how to integrate this type of experience into their own products. Some companies want to make their existing agents more proactive, while others are looking to build them from the ground up.
客户开始询问我们如何将这类体验集成到他们自己的产品中。有些公司希望让现有智能体更具主动性，而另一些公司则希望从头构建它们。
This post breaks down the primitives behind always-on agents and walks through how they fit together.
本文拆解了常驻智能体背后的原语，并逐步说明它们如何组合在一起。
The cycle behind the always-on agent
常驻智能体背后的循环
An always-on agent is not actually “always on.” This is just a description of the user experience.
常驻智能体实际上并不是“始终开启”。这只是对用户体验的一种描述。
In reality, the agent spends most of its time asleep, waking up periodically to do work. Each stretch of work is called a run: the agent receives an event, completes its task, and then saves i...