React logo

React

OrganizationPopular
vercel-labs
react

React renderer for json-render that turns JSON specs into React components. Use when working with @json-render/react, building React UIs from JSON, creating component catalogs, or rendering AI-generated specs.

Overview

Publishervercel-labs
Repositoryjson-render
Skill namereact
Stars
16.5K
Forks
887
Bundled files
Instructions only
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.

  • Self-contained

    Everything the model needs lives in the instructions — no extra files to sync.

  • Open source

    Published by vercel-labs on GitHub. Read the source before you install it.

Installation

Install the React 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.
  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/vercel-labs/json-render.git /tmp/json-render
mkdir -p .claude/skills
cp -r /tmp/json-render/skills/react .claude/skills/react
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable React 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 React 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 React 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.

@json-render/react

React renderer that converts JSON specs into React component trees.

Quick Start

typescript
import { defineRegistry, Renderer } from "@json-render/react";
import { catalog } from "./catalog";

const { registry } = defineRegistry(catalog, {
  components: {
    Card: ({ props, children }) => <div>{props.title}{children}</div>,
  },
});

function App({ spec }) {
  return <Renderer spec={spec} registry={registry} />;
}

Creating a Catalog

typescript
import { defineCatalog } from "@json-render/core";
import { schema } from "@json-render/react/schema";
import { defineRegistry } from "@json-render/react";
import { z } from "zod";

// Create catalog with props schemas
export const catalog = defineCatalog(schema, {
  components: {
    Button: {
      props: z.object({
        label: z.string(),
        variant: z.enum(["primary", "secondary"]).nullable(),
      }),
      description: "Clickable button",
    },
    Card: {
      props: z.object({ title: z.string() }),
      slots: ["default"],
      description: "Card container with title",
    },
    Layout: {
      props: z.object({}),
      slots: ["default", "header", "footer"],
      description: "Layout with named content regions",
    },
  },
});

// Define component implementations with type-safe props
const { registry } = defineRegistry(catalog, {
  components: {
    Button: ({ props }) => (
      <button className={props.variant}>{props.label}</button>
    ),
    Card: ({ props, children }) => (
      <div className="card">
        <h2>{props.title}</h2>
        {children}
      </div>
    ),
    Layout: ({ children, slots }) => (
      <div>
        <header>{slots?.header}</header>
        <main>{children}</main>
        <footer>{slots?.footer}</footer>
      </div>
    ),
  },
});

Spec Structure (Element Tree)

The React schema uses an element tree format:

json
{
  "root": {
    "type": "Card",
    "props": { "title": "Hello" },
    "children": [{ "type": "Button", "props": { "label": "Click me" } }]
  }
}

Named Slots

Use children for the "default" slot. Use the element's top-level slots object for other slot names declared by the catalog:

json
{
  "type": "Layout",
  "props": {},
  "children": ["main"],
  "slots": {
    "header": ["heading"],
    "footer": ["actions"]
  }
}

Registry components receive named content as slots?.header, slots?.footer, and so on. Do not use slots.default.

Visibility Conditions

Use visible on elements to show/hide based on state. New syntax: { "$state": "/path" }, { "$state": "/path", "eq": value }, { "$state": "/path", "not": true }, { "$and": [cond1, cond2] } for AND, { "$or": [cond1, cond2] } for OR. Helpers: visibility.when("/path"), visibility.unless("/path"), visibility.eq("/path", val), visibility.and(cond1, cond2), visibility.or(cond1, cond2).

Providers

ProviderPurpose
StateProviderShare state across components (JSON Pointer paths). Accepts optional store prop for controlled mode.
ActionProviderHandle actions dispatched via the event system
VisibilityProviderEnable conditional rendering based on state
ValidationProviderForm field validation

External Store (Controlled Mode)

Pass a StateStore to StateProvider (or JSONUIProvider / createRenderer) to use external state management (Redux, Zustand, XState, etc.):

tsx
import { createStateStore, type StateStore } from "@json-render/react";

const store = createStateStore({ count: 0 });

<StateProvider store={store}>{children}</StateProvider>;

// Mutate from anywhere — React re-renders automatically:
store.set("/count", 1);

When store is provided, initialState and onStateChange are ignored.

Dynamic Prop Expressions

Any prop value can be a data-driven expression resolved by the renderer before components receive props:

  • { "$state": "/state/key" } - reads from state model (one-way read)
  • { "$bindState": "/path" } - two-way binding: reads from state and enables write-back. Use on the natural value prop (value, checked, pressed, etc.) of form components.
  • { "$bindItem": "field" } - two-way binding to a repeat item field. Use inside repeat scopes.
  • Filtered lists: repeat plus an $item visible condition on the same container renders only matching items: { "repeat": { "statePath": "/tasks", "key": "id" }, "visible": { "$item": "status", "eq": "todo" }, "children": ["task-card"] }. AND-composed $state conjuncts gate the container shell; $item/$index conjuncts filter items.
  • Nested lists: inside a repeat, use { "repeat": { "statePath": { "$item": "comments" }, "key": "id" } } to iterate an array on the enclosing item.
  • { "$cond": <condition>, "$then": <value>, "$else": <value> } - conditional value
  • { "$template": "Hello, ${/name}!" } - interpolates state values into strings
  • { "$computed": "fn", "args": { ... } } - calls registered functions with resolved args
json
{
  "type": "Input",
  "props": {
    "value": { "$bindState": "/form/email" },
    "placeholder": "Email"
  }
}

Components do not use a statePath prop for two-way binding. Use { "$bindState": "/path" } on the natural value prop instead.

Components receive already-resolved props. For two-way bound props, use the useBoundProp hook with the bindings map the renderer provides.

Register $computed functions via the functions prop on JSONUIProvider or createRenderer:

tsx
<JSONUIProvider
  functions={{ fullName: (args) => `${args.first} ${args.last}` }}
>

Event System

Components use emit to fire named events, or on() to get an event handle with metadata. The element's on field maps events to action bindings:

tsx
// Simple event firing
Button: ({ props, emit }) => (
  <button onClick={() => emit("press")}>{props.label}</button>
),

// Event handle with metadata (e.g. preventDefault)
Link: ({ props, on }) => {
  const click = on("click");
  return (
    <a href={props.href} onClick={(e) => {
      if (click.shouldPreventDefault) e.preventDefault();
      click.emit();
    }}>{props.label}</a>
  );
},
json
{
  "type": "Button",
  "props": { "label": "Submit" },
  "on": { "press": { "action": "submit" } }
}

The EventHandle returned by on() has: emit(), shouldPreventDefault (boolean), and bound (boolean).

State Watchers

Elements can declare a watch field (top-level, sibling of type/props/children) to trigger actions when state values change:

json
{
  "type": "Select",
  "props": {
    "value": { "$bindState": "/form/country" },
    "options": ["US", "Canada"]
  },
  "watch": { "/form/country": { "action": "loadCities" } },
  "children": []
}

Built-in Actions

The setState, pushState, removeState, and validateForm actions are built into the React schema and handled automatically by ActionProvider. They are injected into AI prompts without needing to be declared in catalog actions:

json
{ "action": "setState", "params": { "statePath": "/activeTab", "value": "home" } }
{ "action": "pushState", "params": { "statePath": "/items", "value": { "text": "New" } } }
{ "action": "removeState", "params": { "statePath": "/items", "index": 0 } }
{ "action": "validateForm", "params": { "statePath": "/formResult" } }

validateForm validates all registered fields and writes { valid, errors } to state.

Note: statePath in action params (e.g. setState.statePath) targets the mutation path. Two-way binding in component props uses { "$bindState": "/path" } on the value prop, not statePath.

useBoundProp

For form components that need two-way binding, use useBoundProp with the bindings map the renderer provides when a prop uses { "$bindState": "/path" } or { "$bindItem": "field" }:

tsx
import { useBoundProp } from "@json-render/react";

Input: ({ element, bindings }) => {
  const [value, setValue] = useBoundProp<string>(
    element.props.value,
    bindings?.value
  );
  return (
    <input
      value={value ?? ""}
      onChange={(e) => setValue(e.target.value)}
    />
  );
},

useBoundProp(propValue, bindingPath) returns [value, setValue]. The value is the resolved prop; setValue writes back to the bound state path (no-op if not bound).

BaseComponentProps

For building reusable component libraries not tied to a specific catalog (e.g. @json-render/shadcn):

typescript
import type { BaseComponentProps } from "@json-render/react";

const Card = ({ props, children }: BaseComponentProps<{ title?: string }>) => (
  <div>{props.title}{children}</div>
);

defineRegistry

defineRegistry conditionally requires the actions field only when the catalog declares actions. Catalogs with actions: {} can omit it.

Key Exports

ExportPurpose
defineRegistryCreate a type-safe component registry from a catalog
RendererRender a spec using a registry
schemaElement tree schema (includes built-in state actions: setState, pushState, removeState, validateForm)
useStateStoreAccess state context
useStateValueGet single value from state
useBoundPropTwo-way binding for $bindState/$bindItem expressions
useActionsAccess actions context
useActionGet a single action dispatch function
useOptionalValidationNon-throwing variant of useValidation (returns null if no provider)
useUIStreamStream specs from an API endpoint
createStateStoreCreate a framework-agnostic in-memory StateStore
StateStoreInterface for plugging in external state management
BaseComponentPropsCatalog-agnostic base type for reusable component libraries
EventHandleEvent handle type (emit, shouldPreventDefault, bound)
ComponentContextTyped component context (catalog-aware)

Frequently asked questions

What does the React AI skill do?

React renderer for json-render that turns JSON specs into React components. Use when working with @json-render/react, building React UIs from JSON, creating component catalogs, or rendering AI-generated specs.

Why use React on TypingMind?

Because you install it once and use it with any model. React 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 React in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/vercel-labs/json-render/tree/main/skills/react. TypingMind reads its SKILL.md and installs it as a skill you can enable per chat.

Which AI models can use React?

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 React?

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

Is the React AI skill free?

Yes. It is published on GitHub by vercel-labs 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 👇