AI Status Orb new AI 状态球 AI Status Orb web
A compact animated focal point that communicates an AI system's current state
用紧凑的动态焦点表达 AI 系统当前处于什么状态
标本可交互，可以直接操作。
Anatomy — every part, named解剖 —— 每个部件的名字
-
1 Orb core球体核心
isolated visualThe central shape that gives the status indicator a stable visual anchor.
给状态指示器提供稳定视觉锚点的中心形状。
Prompt fragmentPrompt 片段a 64px orb with a quiet center and bounded layered light, never a full-screen glow
64px 状态球，中心安静，分层光照有边界，不做全屏发光
-
2 State motion状态动效
state machineThe distinct motion grammar assigned to each AI state.
分配给每个 AI 状态的不同动效语法。
Prompt fragmentPrompt 片段Idle stays still, Listening breathes slowly, Thinking rotates a restrained halo, Speaking responds to activity, Error holds a clear static warning
Idle 静止，Listening 缓慢呼吸，Thinking 旋转克制光环，Speaking 随活动响应，Error 保持清晰静态警示
-
3 State label状态标签
role=statusThe words that make the visual state understandable and searchable.
让视觉状态可理解、可搜索的文字。
Prompt fragmentPrompt 片段a persistent text label such as Listening or Thinking beside the orb, announced politely
球旁持续显示 Listening 或 Thinking 等文字标签，并礼貌播报
Prompt — paste into your agentPrompt —— 直接粘贴给你的代理
Build an AI status orb with explicit Idle, Listening, Thinking, Speaking, and Error states. Pair every visual state with a visible text label and role=status announcement, use a bounded layered glow and slow scale/pulse only for active states, stop or pause animation when hidden, keep the control keyboard accessible, and freeze the orb with a static state marker under prefers-reduced-motion.
构建一个 AI 状态球，包含 Idle、Listening、Thinking、Speaking 和 Error 五种明确状态。每个视觉状态都配有可见文字标签和 role=status 播报；活动状态只使用有边界的分层光晕和缓慢缩放/脉冲；隐藏时停止或暂停动画；控件支持键盘；在 prefers-reduced-motion 下冻结球体并保留静态状态标记。
Debug prompt — when it misbehaves调试 Prompt —— 当它不听话时
Debug my AI status orb. Check that the animation is driven by an explicit state machine rather than scattered booleans; Idle is visually distinct from Thinking and Speaking; the state is announced in text and aria-live without repeating every frame; focus and pointer controls remain usable; offscreen and hidden states stop expensive work; errors have a recovery action; and reduced-motion freezes movement while preserving the state distinction. The symptom:
调试我的 AI 状态球。检查：动画是否由明确状态机驱动，而不是散落的布尔值；Idle 是否和 Thinking、Speaking 有区别；状态是否通过文字和 aria-live 播报且不会重复每一帧；焦点和指针控件是否可用；离屏和隐藏状态是否停止昂贵工作；错误是否有恢复操作；reduced-motion 是否冻结运动但保留状态区别。症状：
In code代码里叫什么
| Framework框架 | Symbol符号 | Note说明 |
|---|---|---|
| ARIA | role=status + aria-live | name the state in words beside the visual在视觉旁用文字说明状态 |
| Motion | state machine | map idle, listening, thinking, speaking, and error to deliberate motion把 idle、listening、thinking、speaking 和 error 映射到明确动效 |
| CSS | radial-gradient + transform | bound the glow and animate only the isolated focal region限制光晕范围，只在独立焦点区域做动画 |
| Motion | prefers-reduced-motion | freeze the pulse and retain a clear state marker冻结脉冲，但保留清晰的状态标记 |