Skip to content

🎉 Material UI v5 is out! Head to the migration guide to get started.


The autocomplete is a normal text input enhanced by a panel of suggested options.

The widget is useful for setting the value of a single-line textbox in one of two types of scenarios:

  1. The value for the textbox must be chosen from a predefined set of allowed values, e.g., a location field must contain a valid location name: combo box.
  2. The textbox may contain any arbitrary value, but it is advantageous to suggest possible values to the user, e.g., a search field may suggest similar or previous searches to save the user time: free solo.

It's meant to be an improved version of the "react-select" and "downshift" packages.

Combo box

The value must be chosen from a predefined set of allowed values.

  getOptionLabel={(option) => option.title}
  style={{ width: 300 }}
  renderInput={(params) => <TextField {...params} label="Combo box" variant="outlined" />}


Each of the following examples demonstrate one feature of the Autocomplete component.

Country select

Choose one of the 248 countries.

Controllable states

The component has two states that can be controlled:

  1. the "value" state with the value/onChange props combination. This state represents the value selected by the user, for instance when pressing Enter.
  2. the "input value" state with the inputValue/onInputChange props combination. This state represents the value displayed in the textbox.

⚠️ These two state are isolated, they should be controlled independently.

value: 'Option 1'
inputValue: ''

Free solo

Set freeSolo to true so the textbox can contain any arbitrary value.

Search input

The prop is designed to cover the primary use case of a search input with suggestions, e.g. Google search or react-autowhatever.


If you intend to use this mode for a combo box like experience (an enhanced version of a select element) we recommend setting:

  • selectOnFocus to helps the user clear the selected value.
  • clearOnBlur to helps the user to enter a new value.
  • handleHomeEndKeys to move focus inside the popup with the Home and End keys.
  • A last option, for instance Add "YOUR SEARCH".

You could also display a dialog when the user wants to add a new value.


  options={options.sort((a, b) => -b.firstLetter.localeCompare(a.firstLetter))}
  groupBy={(option) => option.firstLetter}
  getOptionLabel={(option) => option.title}
  style={{ width: 300 }}
  renderInput={(params) => <TextField {...params} label="With categories" variant="outlined" />}

Disabled options

  getOptionDisabled={(option) => option === timeSlots[0] || option === timeSlots[2]}
  style={{ width: 300 }}
  renderInput={(params) => (
    <TextField {...params} label="Disabled options" variant="outlined" />


For advanced customization use cases, we expose a headless useAutocomplete() hook. It accepts almost the same options as the Autocomplete component minus all the props related to the rendering of JSX. The Autocomplete component uses this hook internally.

import useAutocomplete from '@material-ui/lab/useAutocomplete';

Customized hook

Head to the Customized Autocomplete section for a customization example with the Autocomplete component instead of the hook.

Asynchronous requests

Google Maps place

A customized UI for Google Maps Places Autocomplete.

For this demo, we need to load the Google Maps JavaScript API.

⚠️ Before you can start using the Google Maps JavaScript API, you must sign up and create a billing account.

Multiple values

Also known as tags, the user is allowed to enter more than one value.

Fixed options

In the event that you need to lock certain tag so that they can't be removed in the interface, you can set the chips disabled.


Limit tags

You can use the limitTags prop to limit the number of displayed options when not focused.

  getOptionLabel={(option) => option.title}
  defaultValue={[top100Films[13], top100Films[12], top100Films[11]]}
  renderInput={(params) => (
    <TextField {...params} variant="outlined" label="limitTags" placeholder="Favorites" />


Fancy smaller inputs? Use the size prop.


Custom input

The renderInput prop allows you to customize the rendered input. The first argument of this render prop contains props that you need to forward. Pay specific attention to the ref and inputProps keys.

  renderInput={(params) => (
    <div ref={params.InputProps.ref}>
      <input style={{ width: 200 }} type="text" {...params.inputProps} />

GitHub's picker

This demo reproduces the GitHub's label picker:

help wanted
type: bug

Head to the Customized hook section for a customization example with the useAutocomplete hook instead of the component.


The following demo relies on autosuggest-highlight, a small (1 kB) utility for highlighting text in autosuggest and autocomplete components.

Custom filter

The component exposes a factory to create a filter method that can provided to the filterOptions prop. You can use it to change the default option filter behavior.

import { createFilterOptions } from '@material-ui/lab/Autocomplete';

createFilterOptions(config) => filterOptions


  1. config (Object [optional]):
    • config.ignoreAccents (Boolean [optional]): Defaults to true. Remove diacritics.
    • config.ignoreCase (Boolean [optional]): Defaults to true. Lowercase everything.
    • config.limit (Number [optional]): Default to null. Limit the number of suggested options to be shown. For example, if config.limit is 100, only the first 100 matching options are shown. It can be useful if a lot of options match and virtualization wasn't set up.
    • config.matchFrom ('any' | 'start' [optional]): Defaults to 'any'.
    • config.stringify (Func [optional]): Controls how an option is converted into a string so that it can be matched against the input text fragment.
    • config.trim (Boolean [optional]): Defaults to false. Remove trailing spaces.


filterOptions: the returned filter method can be provided directly to the filterOptions prop of the Autocomplete component, or the parameter of the same name for the hook.

In the following demo, the options need to start with the query prefix:

const filterOptions = createFilterOptions({
  matchFrom: 'start',
  stringify: option => option.title,

<Autocomplete filterOptions={filterOptions} />


For richer filtering mechanisms, like fuzzy matching, it's recommended to look at match-sorter. For instance:

import matchSorter from 'match-sorter';

const filterOptions = (options, { inputValue }) =>
  matchSorter(options, inputValue);

<Autocomplete filterOptions={filterOptions} />


Search within 10,000 randomly generated options. The list is virtualized thanks to react-window.

  style={{ width: 300 }}
  groupBy={(option) => option[0].toUpperCase()}
  renderInput={(params) => <TextField {...params} variant="outlined" label="10,000 options" />}
  renderOption={(option) => <Typography noWrap>{option}</Typography>}



The browsers have heuristics to help the users fill the form inputs. However, it can harm the UX of the component.

By default, the component disable the autocomplete feature (remembering what the user has typed for a given field in a previous session) with the autoComplete="off" attribute.

However, in addition to remembering past entered values, the browser might also propose autofill suggestions (saved login, address, or payment details). In the event you want the avoid autofill, you can try the following:

  • Name the input without leaking any information the browser can use. e.g. id="field1" instead of id="country". If you leave the id empty, the component uses a random id.
  • Set autoComplete="new-password":
        autoComplete: 'new-password',

iOS VoiceOver

VoiceOver on iOS Safari doesn't support the aria-owns attribute very well. You can work around the issue with the disablePortal prop.


If you provide a custom ListboxComponent prop, you need to make sure that the intended scroll container has the role attribute set to listbox. This ensures the correct behavior of the scroll, for example when using the keyboard to navigate.



We encourage the usage of a label for the textbox. The component implements the WAI-ARIA authoring practices.