Thinking Orbs logo

Thinking Orbs

Community
boraoztunc
thinking-orbs

Add accessible animated AI loading and agent-status indicators with the React thinking-orbs library. Use when a chat, copilot, voice, search, generation, or tool-running interface needs a semantic working, searching, solving, listening, composing, or shaping state; when replacing a generic spinner with an AI activity orb; or when implementing the library's size, theme, speed, pause, reduced-motion, and canvas behavior.

Overview

Publisherboraoztunc
Repositoryskills
Skill namethinking-orbs
Stars
422
Forks
63
Bundled files
1
LicenseApache-2.0
Links
  • Markdown instructions

    A SKILL.md file the model loads on demand, so it only costs tokens when a request actually matches.

  • Works with any LLM

    AI skills are plain Markdown, not provider-specific code, so this works with GPT, Claude, Gemini, Grok, or a local model.

  • 1 bundled files

    Scripts, templates, and references the model can read while it works. Files are read-only and never executed.

  • Open source

    Published by boraoztunc on GitHub. Read the source before you install it.

Installation

Install the Thinking Orbs AI skill in TypingMind to use it with any LLM, or drop it into another agent that reads SKILL.md.

1

Install in TypingMind

TypingMind installs a skill straight from its GitHub folder — it reads SKILL.md, bundles the resource files, and stores the result locally.

  1. Open the app and go to Plugins → Skills.
  2. Choose "Install from GitHub".
  3. Paste the skill folder URL below and confirm.
    https://github.com/boraoztunc/skills/tree/main/thinking-orbs
  4. Enable the skill in any chat where you want it available.
Plugins → Skills → Add skill → From GitHub URL, then paste the folder URL and press Continue.
2

Install in another agent

Any agent that reads the Agent Skills format can use this skill — copy the folder into that agent's skills directory.

Claude Code — .claude/skills
git clone --depth 1 https://github.com/boraoztunc/skills.git /tmp/skills
mkdir -p .claude/skills
cp -r /tmp/skills/thinking-orbs .claude/skills/thinking-orbs
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Thinking Orbs in any TypingMind chat and the model takes it from there. Its name and description sit in the system prompt, and the moment a request matches, the model loads the full instructions itself — you never invoke it by hand, and it costs no tokens until it is actually used.

The model loads Thinking Orbs on its own as soon as a request matches it.

Works with any AI model

AI skills are plain Markdown instructions rather than provider-specific code, so Thinking Orbs is not tied to the model it was written for. Install it once in TypingMind and use it with GPT-5, Claude, Gemini, Grok, DeepSeek, Mistral, Llama, or a local model you run yourself — all on your own API keys.

  • Loaded only when it is needed

    The system prompt carries just the name and description. The instructions are fetched on the first matching request, so an idle skill costs nothing.

  • Switch models mid-chat

    Because the skill is instructions rather than code, changing model does not break it — the next model reads the same SKILL.md.

Skill instructions

This is the SKILL.md content the model loads. Read it before installing — a skill is instructions your model will follow.

Thinking Orbs

Core Contract

  1. Use thinking-orbs in React 18+ interfaces that need indeterminate AI activity feedback.
  2. Map the real product lifecycle to one of the six shipped states.
  3. Use only the tuned 20 or 64 pixel size. Do not stretch one preset into another.
  4. Keep theme="auto" unless the surrounding surface has a known fixed theme.
  5. Pair the orb with concise visible status text when the activity matters to the user.
  6. Override aria-label with a task-specific label, or hide the orb from assistive technology when adjacent live text already announces the same state.
  7. Use paused to freeze the current frame. Do not simulate pause with speed={0}.
  8. Treat the orb as indeterminate feedback, never as a progress percentage or completion signal.

The package renders monochrome dots on a transparent 2D canvas. It does not expose custom colors, arbitrary sizes, or determinate progress.

Install

Inspect the project package manager, then install:

bash
npm install thinking-orbs

The package declares react and react-dom version 18 or newer as peer dependencies. Import the component and exported types from the package root:

tsx
import {
  ThinkingOrb,
  type OrbSize,
  type OrbState,
  type OrbTheme,
  type ThinkingOrbProps,
} from "thinking-orbs";

Choose the State

  • working — generic tool execution, multi-step work, or an activity without a more precise state.
  • searching — retrieval, web search, file search, or knowledge lookup.
  • solving — reasoning, analysis, calculation, or planning.
  • listening — microphone input, speech capture, or waiting for a spoken turn.
  • composing — writing, summarizing, drafting, or generating a text response.
  • shaping — creating or refining an image, layout, structured artifact, or other formed output.

Prefer a truthful generic working state over a visually interesting but inaccurate state. Change the state only when the underlying activity changes.

Choose the Size

  • Use size={20} inline with text, inside buttons, or in compact status rows.
  • Use size={64} at chat-avatar scale, in an empty state, or as the main visual status.

The presets have different dot counts, dot sizes, and speed tuning. They are separate designs rather than a scale factor. If the layout needs more surrounding space, size the wrapper instead of applying CSS transforms to the canvas.

Basic Usage

tsx
import { ThinkingOrb } from "thinking-orbs";

export function AgentStatus() {
  return (
    <ThinkingOrb
      state="searching"
      size={20}
      theme="auto"
      aria-label="Searching project files…"
    />
  );
}

All other canvas props pass through, including className, style, data-*, event handlers, and ARIA attributes.

Model the Product Lifecycle

Keep product phases separate from visual states so the mapping stays explicit:

tsx
import { ThinkingOrb, type OrbState } from "thinking-orbs";

type AgentPhase =
  | "idle"
  | "retrieving"
  | "reasoning"
  | "writing"
  | "creating"
  | "done"
  | "error";

const ORB_BY_PHASE: Partial<Record<AgentPhase, OrbState>> = {
  retrieving: "searching",
  reasoning: "solving",
  writing: "composing",
  creating: "shaping",
};

export function AgentActivity({ phase }: { phase: AgentPhase }) {
  const state = ORB_BY_PHASE[phase];
  if (!state) return null;

  return <ThinkingOrb state={state} size={20} />;
}

Remove the orb on done, error, cancellation, or idle. Show the appropriate result, retry, or error UI instead of leaving the last activity animation running.

Announce Status Once

The component defaults to role="img" with a per-state label such as “Searching…”. When visible text describes the same state, make the text the single announcement source:

tsx
export function LiveAgentStatus() {
  return (
    <div role="status" aria-live="polite" className="agent-status">
      <ThinkingOrb state="solving" size={20} aria-hidden="true" />
      <span>Reviewing the repository…</span>
    </div>
  );
}

Use aria-live="polite" for ordinary phase changes. Avoid rapid label churn. Do not add another hidden live region when role="status" already owns the announcement.

Theme

Use one of:

tsx
<ThinkingOrb theme="auto" />
<ThinkingOrb theme="dark" />
<ThinkingOrb theme="light" />
  • auto first checks an ancestor data-theme="dark|light" attribute or dark / light class.
  • If no ancestor theme exists, auto follows prefers-color-scheme.
  • Theme changes update live.
  • dark means light dots intended for a dark background.
  • light means dark dots intended for a light background.

The canvas is transparent. Verify contrast against the actual surface rather than the page root alone.

Speed and Pause

tsx
<ThinkingOrb state="working" speed={0.85} />
<ThinkingOrb state="composing" paused={isWaitingForApproval} />

speed multiplies the baked speed of the selected state and size. Start at 1; use roughly 0.75–1.25 for subtle product tuning. Extreme values can make the hand-tuned motion feel frantic or stalled.

paused freezes the current frame while retaining the visual status. Remove the component when the activity has actually ended.

Next.js and Client Rendering

The component uses React effects, canvas, requestAnimationFrame, media queries, and observers. Keep the package import behind a client boundary in the Next.js App Router:

tsx
"use client";

import { ThinkingOrb } from "thinking-orbs";

export function ThinkingStatus() {
  return <ThinkingOrb state="working" size={20} />;
}

The library is SSR-safe because it paints only on the client after resolving the theme. A client boundary is still required where the framework enforces server and client component separation.

Built-In Runtime Behavior

  • Draws with plain Canvas 2D arcs; no WebGL or SVG filters.
  • Caps device pixel ratio at 2.
  • Uses one requestAnimationFrame loop per visible instance.
  • Uses the shared performance.now() clock so multiple orbs stay in phase.
  • Pauses when the canvas scrolls offscreen through IntersectionObserver.
  • Pauses when the browser tab is hidden.
  • Renders one deterministic static frame under prefers-reduced-motion: reduce.
  • Continues following live theme changes in reduced-motion mode.

Do not rebuild these behaviors in a wrapper. Add product state management and layout around the component, not a second animation loop.

Power-User Canvas API

Prefer <ThinkingOrb> for product UI. The package also exports its resolved presets and raw frame painters for a custom canvas outside React:

ts
import { MODE_DRAWS, resolvePreset } from "thinking-orbs";

const { mode, speed, opts } = resolvePreset("searching", 64);
const drawFrame = MODE_DRAWS[mode];

drawFrame(
  context,
  64,
  (performance.now() / 1000) * speed,
  true, // true draws light ink for a dark surface
  opts,
);

STATE_TO_MODE exposes the internal mapping:

  • working → orbits
  • searching → globe
  • solving → rubik
  • listening → wave
  • composing → ribbon
  • shaping → morph

Use the raw API only when another renderer owns the canvas lifecycle. It provides a frame painter, not component behavior. Reimplement DPR sizing, clearing, animation scheduling, pausing, theme resolution, reduced motion, visibility handling, cleanup, and accessibility when bypassing <ThinkingOrb>.

Verification

Run the project's typecheck, tests, production build, and git diff --check. Then verify in a real browser:

  1. Trigger every product phase and confirm the mapped orb state is truthful.
  2. Confirm 20 and 64 pixel instances are crisp without CSS scaling.
  3. Test dark, light, and live theme switching.
  4. Test reduced motion and confirm the orb becomes a static representative frame.
  5. Scroll the orb offscreen and return; confirm it resumes without visible breakage.
  6. Hide and restore the tab; confirm animation resumes.
  7. Inspect accessibility: announce the status exactly once and use a task-specific label.
  8. Confirm the orb disappears on success, error, cancellation, and idle.
  9. Confirm no console errors, hydration warnings, or layout shifts occur.

Common Pitfalls

  • Type error on size: use exactly 20 or 64; do not pass arbitrary dimensions.
  • Wrong contrast: remember dark targets dark backgrounds and therefore draws light ink.
  • Duplicate screen-reader output: hide the canvas when adjacent role="status" text already announces the task.
  • Misleading state: do not show searching during generation or composing during microphone capture.
  • Permanent loading UI: remove the orb when work ends and render the actual terminal state.
  • Hydration or server-component error: move the import into a client component.
  • Brand color requested: the public API is monochrome; choose another loader or make an intentional library fork instead of relying on unsupported styling.

Handoff

Report the mapped product phases, chosen size, theme mode, accessible label strategy, reduced-motion behavior, and build/browser verification. Distinguish local implementation from a deployed release.

Bundled files

The model reads these on demand while the skill is loaded. They are exposed as readable files and are never executed.

Frequently asked questions

What does the Thinking Orbs AI skill do?

Add accessible animated AI loading and agent-status indicators with the React thinking-orbs library. Use when a chat, copilot, voice, search, generation, or tool-running interface needs a semantic working, searching, solving, listening, composing, or shaping state; when replacing a generic spinner with an AI activity orb; or when implementing the library's size, theme, speed, pause, reduced-motion, and canvas behavior.

Why use Thinking Orbs on TypingMind?

Because you install it once and use it with any model. Thinking Orbs is plain Markdown rather than provider-specific code, so the same skill runs on GPT-5, Claude, Gemini, Grok, or a local model — and you can switch model mid-chat without it breaking. TypingMind runs on your own API keys, so you pay providers directly instead of a per-seat subscription, and your skills and chats stay in your own storage.

How do I install Thinking Orbs in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/boraoztunc/skills/tree/main/thinking-orbs. TypingMind reads its SKILL.md and bundles its files and installs it as a skill you can enable per chat.

Which AI models can use Thinking Orbs?

Any model you connect in TypingMind. AI skills are plain Markdown instructions rather than provider-specific code, so GPT, Claude, Gemini, Grok, and local models can all load this skill when a request matches it.

How many AI models can I use with Thinking Orbs?

As many as you like. As long as a model supports skills, you can use Thinking Orbs with it — GPT, Claude, Gemini, Grok, DeepSeek, Mistral, Llama and more — all on TypingMind with your own API keys.

Is the Thinking Orbs AI skill free?

Yes. It is published on GitHub by boraoztunc under the Apache-2.0 license. You only pay your own AI provider for the tokens you use.

What are AI skills?

An AI skill is a reusable instruction bundle that teaches an AI model how to do one specific task. It follows the open Agent Skills format: a SKILL.md file with a name and description, plus any scripts, templates or reference files the model may need. The model reads the instructions only when your request matches the skill, so an installed skill costs nothing until it is used.

How are AI skills different from plugins or MCP servers?

A plugin or MCP server gives a model new tools to call — code that runs somewhere and returns a result. An AI skill gives the model knowledge and process instead: how to approach a task, which steps to follow, what good output looks like. Skills are plain Markdown, so they need no server, no API key and no runtime, and they work with any model.

View all

Set up your own AI workspace now

Get notified about new features and future giveaways by subscribing to our newsletter 👇