Lynx Ui Input logo

Lynx Ui Input

Organization
lynx-family
lynx-ui-input

Build and troubleshoot text entry with lynx-ui Input, TextArea, and KeyboardAware primitives, including controlled values, native filtering, selection, and keyboard avoidance.

Overview

Publisherlynx-family
Repositorylynx-ui
Skill namelynx-ui-input
Stars
119
Forks
13
Bundled files
20
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.

  • 20 bundled files

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

  • Open source

    Published by lynx-family on GitHub. Read the source before you install it.

Installation

Install the Lynx Ui Input 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/lynx-family/lynx-ui.git /tmp/lynx-ui
mkdir -p .claude/skills
cp -r /tmp/lynx-ui/packages/lynx-ui-input .claude/skills/lynx-ui-input
Restart Claude Code after copying so it picks up the new skill.

Use it in TypingMind

Enable Lynx Ui Input 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 Lynx Ui Input 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 Lynx Ui Input 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.

lynx-ui-input SKILL

lynx-ui-input provides headless single-line and multiline text entry for ReactLynx, plus composable keyboard avoidance. Import its public components and ref types from @lynx-js/lynx-ui.

1. Core Capabilities

  • Text entry: Use Input for single-line fields and TextArea for multiline content. Both accept className and style without imposing a visual theme.
  • Value ownership: Use value + onInput for controlled fields, or defaultValue for an initial uncontrolled value.
  • Native editing: Configure keyboard type, confirmation key, character filtering, length limits, and read-only behavior through the wrapper props.
  • Editing callbacks: Receive text, selection offsets, and optional input method editor (IME) composition state without unpacking native event objects.
  • Imperative operations: InputRef and TextAreaRef expose asynchronous focus, blur, value access, and selection methods.
  • Keyboard avoidance: Compose KeyboardAwareRoot, KeyboardAwareResponder, and KeyboardAwareTrigger to move a region or scroll the focused field into view.

2. AI Coding Guide

Minimal Usable Example (Controlled)

tsx
import { useState } from '@lynx-js/react'
import { Input, TextArea } from '@lynx-js/lynx-ui'

function ProfileFields() {
  const [name, setName] = useState('')
  const [bio, setBio] = useState('')

  return (
    <view>
      <Input
        value={name}
        onInput={setName}
        placeholder='Name'
        maxLength={60}
        style={{ width: '100%', height: '44px' }}
      />
      <TextArea
        value={bio}
        onInput={setBio}
        placeholder='About you'
        maxLength={600}
        maxLines={8}
        style={{ width: '100%', height: '120px' }}
      />
    </view>
  )
}

Give fields a usable width and a deliberate height or content-sizing policy. Compose labels, borders, icons, clear buttons, and validation messages in surrounding views.

Public API Boundary

The wrappers render <input> and <textarea>. The native x-input-ng and x-textarea-ng implementations inform the behavior notes below, but their complete native attribute and event surface is not the ReactLynx wrapper API.

Wrapper APINative mapping / behavior
value, defaultValueApplied through the native setValue method.
onInput(value, start, end, isComposing?)Unpacks bindinput details into positional arguments.
onFocus(value), onBlur(value), onConfirm(value)Each receives the text string.
onSelectionChange(start, end)Unpacks bindselection; offsets are native text positions.
maxLengthmaxlength; wrapper default is 140.
inputFilterinput-filter; pass a regex string.
confirmTypeconfirm-type; wrapper default is 'send'.
showSoftInputOnFocusshow-soft-input-on-focus; defaults to true.
TextArea.maxLines, lineSpacingmaxlines (default 40) and line-spacing.
TextArea.bouncesiOS bounce behavior; defaults to true.

Use readonly (lowercase), not DOM readOnly or disabled. These wrappers do not expose onChange, autoFocus, rows, autoHeight, raw inputProps / textareaProps, or arbitrary native prop spreading. Native options such as confirm-enter and native line-count or keyboard-height events are not exposed either. Do not add these props to generated consumer code.

Recommended Prompt Formula

Field: [single-line Input / multiline TextArea] for [purpose]. Value ownership: [controlled value + onInput / uncontrolled defaultValue]. Editing rules: [keyboard type], [allowed characters], [length / line limit], and [validation timing]. Keyboard behavior: [confirmation action], [focus or selection action], and [no avoidance / translated region / scrollable form]. Appearance: [dimensions], [labels and surrounding controls].

Example Prompts:

  • "Build a controlled search Input with confirmType='search'. Update the text immediately and debounce search requests while respecting IME composition."
  • "Build an uncontrolled numeric Input with a six-character limit and a clear button using InputRef.setValue('')."
  • "Build a profile form with Input and TextArea inside a vertical KeyboardAwareResponder. Keep each focused field above the keyboard."

3. Use Cases & Best Practices

Value Synchronization and IME Composition

  • Choose the mode at mount and keep it stable. The wrappers capture value !== undefined once; initialize controlled state to '', not undefined. Do not combine value and defaultValue.
  • Mirror the incoming text in onInput immediately, including composing text. Debounce requests or expensive validation separately from the value update. Controlled input briefly sets native readonly while dispatching the callback to the background thread, then unlocks it when the callback returns; keep this callback short and avoid throwing from it.
  • Skip destructive formatting, cursor rewrites, and submission while isComposing === true. Keep the raw editing value separate from a normalized submitted value when necessary.
  • Do not rely on returning the same controlled value to reject a native edit: the value effect runs when the prop changes. Prefer inputFilter for simple character restrictions and validate complete values on blur or confirmation.
  • defaultValue is initialization only. Reset uncontrolled fields through ref.setValue(''); reset controlled fields through their state setter.
  • Do not assume programmatic setValue is silent or produces exactly one onInput. Android text replacement can notify text watchers, and iOS TextArea explicitly runs its change handler. Keep callbacks safe for repeated values.

Filtering and Multiline Input

type accepts 'text', 'number', 'digit', 'password', 'tel', and 'email'. It selects native editing / keyboard behavior; validate complete business values separately. For a simple ASCII digit field:

tsx
<Input
  defaultValue=''
  type='number'
  inputFilter='[0-9]'
  maxLength={6}
  style={{ width: '100%', height: '44px' }}
/>

Android and iOS native filters inspect individual characters. Use allowlists such as [0-9] or [A-Za-z0-9], not a whole-value regex such as ^\d{6}$ or a complete email pattern. For unrestricted human-language text, avoid character filters that could discard composing text. Harmony delegates filtering to ArkUI, so verify paste and IME behavior on the target platform.

Set maxLength explicitly for long-form text; the wrapper's default remains 140 even for TextArea. maxLines is a native line constraint, not a CSS height or an HTML rows equivalent. Android measures wrapped text when enforcing the limit, while other platforms use different native layout paths. Test wrapping, paste, line breaks, and font sizing when the exact limit matters.

Use lineSpacing for additional line spacing and bounces for the iOS scrolling effect. Do not infer portable multiline password support from the inherited type union: the inspected Harmony TextArea parser does not support 'password'. Use Input type='password' for password entry.

Focus, Value Access, and Selection

All five ref methods return promises. Call them after the component mounts, guard the ref, and handle native invocation failures. Native failures are wrapped in InvokeRejectError.

tsx
import { useRef } from '@lynx-js/react'
import { Input } from '@lynx-js/lynx-ui'
import type { InputRef } from '@lynx-js/lynx-ui'

function EditableName() {
  const inputRef = useRef<InputRef>(null)

  const selectAll = async () => {
    const input = inputRef.current
    if (!input) return

    try {
      await input.focus()
      const { value } = await input.getValue()
      await input.setSelectionRange(0, value.length)
    } catch (error) {
      console.error('Could not select the input text', error)
    }
  }

  return (
    <view>
      <Input
        ref={inputRef}
        defaultValue='Alex'
        style={{ width: '100%', height: '44px' }}
      />
      <view bindtap={selectAll}>
        <text>Select all</text>
      </view>
    </view>
  )
}
  • getValue() resolves to { value, selectionStart, selectionEnd } in the public type. The wrapper does not type the native isComposing result here; consume composition state through onInput.
  • Selection offsets can be -1 while unfocused. Focus before restoring a selection and keep 0 <= start <= end <= value.length. Native offsets are not counts of visible glyphs; do not split emoji or composed characters when calculating a partial range.
  • setSelectionRange(position, position) moves the caret. setValue accepts only a string; native { value, cursor } parameters are not its public signature. Await value updates before applying a dependent selection, and read back the text if native filtering may have changed its length.
  • Use blur() to explicitly release focus. showSoftInputOnFocus={false} suppresses the soft keyboard; it does not make the field read-only.

Keyboard-Aware Forms

Android prerequisite: When using keyboard avoidance, the hosting Activity must set soft_input_mode to nothing. This corresponds to android:windowSoftInputMode="adjustNothing" in the Activity manifest or WindowManager.LayoutParams.SOFT_INPUT_ADJUST_NOTHING in native window configuration. System resize or pan behavior can otherwise interfere with keyboard avoidance. Configure this in the Android host; it is not an Input, TextArea, or KeyboardAwareRoot prop.

tsx
import {
  Input,
  KeyboardAwareResponder,
  KeyboardAwareRoot,
  KeyboardAwareTrigger,
  TextArea,
} from '@lynx-js/lynx-ui'

function KeyboardAwareForm() {
  return (
    <KeyboardAwareRoot>
      <KeyboardAwareResponder
        as='ScrollView'
        scrollviewId='profile-fields'
        style={{ width: '100%', height: '100%' }}
      >
        <KeyboardAwareTrigger>
          <Input
            placeholder='Name'
            style={{ width: '100%', height: '44px' }}
          />
        </KeyboardAwareTrigger>
        <KeyboardAwareTrigger>
          <TextArea
            placeholder='About you'
            maxLength={600}
            style={{ width: '100%', height: '120px' }}
          />
        </KeyboardAwareTrigger>
      </KeyboardAwareResponder>
    </KeyboardAwareRoot>
  )
}
  • Place all three primitives under the same root. Wrap each field, or the field plus the controls that must remain visible, in a trigger. Avoid one large trigger around an entire form: its bottom edge becomes the avoidance target.
  • Use one responder per root; the root stores one responder ref. The root is a context provider, so apply dimensions and theme classes to actual views.
  • as='View' is the default and translates the responder. The root writes its transform and transition; put independent animations on a nested view. forceAttach applies to this translation path, aligning the trigger to the keyboard even when it would otherwise remain unobscured.
  • as='ScrollView' creates a vertical ScrollView, adds a keyboard spacer, and scrolls the focused trigger into view. Give its parent a bounded height and supply a unique scrollviewId when multiple scroll views exist.
  • offset is a signed adjustment in Lynx px: use a negative value to leave extra space above the keyboard. It is not a positive gap size.
  • On Android, supply the host's actual androidStatusBarPlusBottomBarHeight in Lynx px. For an immersive status bar, include only the bottom navigation bar. Do not copy a demo device's constant into production.
  • Avoidance depends on the host's global keyboardstatuschanged event. The wrappers set ignore-focus, but do not expose native avoidance options; check host keyboard configuration if nothing moves or movement is duplicated.

4. FAQ

Q: Why does onInput code using event.target.value fail?

A: The first argument is already a string. Use onInput={setValue} or onInput={(value, start, end, isComposing) => { /* ... */ }}.

Q: Does isComposing === false guarantee composition has finished everywhere?

A: No. The inspected Android and iOS implementations inspect native composing text, but Harmony's UIBaseInput::SendInputEvent reports false unconditionally. The public callback also marks this argument optional. Preserve raw text and verify composition-sensitive behavior on the deployed runtime; do not treat this flag as a universal commit signal.

Q: Does confirmType='next' move focus, or does 'send' guarantee submission?

A: It configures the native confirmation key. Handle submission in onConfirm and explicitly focus the next field through its ref when required. Native confirmation, newline insertion, and blur behavior vary by platform, especially for TextArea; confirm-enter is not exposed by the wrapper. Use a separate submit control when a multiline editor needs an unambiguous submit action.

Q: Where are the clear button, error state, and ui-focused styles?

A: Input and TextArea expose neither children/render props nor typed ui-* variants in the current implementation. Compose controls and messages around the field, track focus with onFocus / onBlur, and style consumer-owned views. Do not assume the slot or state-variant APIs from InputOTP apply to these components.

5. Sub Components

  • Input: Single-line native editor; exposes InputRef.
  • TextArea: Multiline native editor with line and bounce options; TextAreaRef has the same methods as InputRef.
  • KeyboardAwareRoot: Owns keyboard height, focused trigger, and responder coordination without rendering a layout node.
  • KeyboardAwareTrigger: Measures the region to keep visible and receives descendant Input/TextArea focus and blur notifications through context.
  • KeyboardAwareResponder: Moves a view or manages a vertical scroll region in response to the focused trigger and keyboard height.

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 Lynx Ui Input AI skill do?

Build and troubleshoot text entry with lynx-ui Input, TextArea, and KeyboardAware primitives, including controlled values, native filtering, selection, and keyboard avoidance.

Why use Lynx Ui Input on TypingMind?

Because you install it once and use it with any model. Lynx Ui Input 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 Lynx Ui Input in TypingMind?

Open Plugins → Skills → Install from GitHub in TypingMind and paste https://github.com/lynx-family/lynx-ui/tree/main/packages/lynx-ui-input. 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 Lynx Ui Input?

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 Lynx Ui Input?

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

Is the Lynx Ui Input AI skill free?

Yes. It is published on GitHub by lynx-family 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 👇