RecognizerIntent

A broadcast intent which can be fired to the BroadcastReceiver component specified in the meta-data defined in the DETAILS_META_DATA meta-data of an Activity satisfying ACTION_WEB_SEARCH.

When fired with sendOrderedBroadcast(Intent, String, BroadcastReceiver, android.os.Handler, int, String, android.os.Bundle), a Bundle of extras will be returned to the provided result receiver, and should ideally contain values for EXTRA_LANGUAGE_PREFERENCE and EXTRA_SUPPORTED_LANGUAGES.

(Whether these are actually provided is up to the particular implementation. It is recommended that Activitys implementing ACTION_WEB_SEARCH provide this information, but it is not required.)

Starts an activity that will prompt the user for speech and send it through a speech recognizer. The results will be returned via activity results (in onActivityResult(int, int, Intent), if you start the intent using startActivityForResult(Intent, int)), or forwarded via a PendingIntent if one is provided.

Starting this intent with just startActivity(Intent) is not supported. You must either use startActivityForResult(Intent, int), or provide a PendingIntent, to receive recognition results.

The implementation of this API is likely to stream audio to remote servers to perform speech recognition which can use a substantial amount of bandwidth.

Required extras:

• EXTRA_LANGUAGE_MODEL

Optional extras:

• EXTRA_PROMPT
• EXTRA_LANGUAGE
• EXTRA_MAX_RESULTS
• EXTRA_RESULTS_PENDINGINTENT
• EXTRA_RESULTS_PENDINGINTENT_BUNDLE

Result extras (returned in the result, not to be specified in the request):

• EXTRA_RESULTS

NOTE: There may not be any applications installed to handle this action, so you should make sure to catch ActivityNotFoundException.

Starts an activity that will prompt the user for speech without requiring the user's visual attention or touch input. It will send it through a speech recognizer, and either synthesize speech for a web search result or trigger another type of action based on the user's speech. This activity may be launched while device is locked in a secure mode. Special care must be taken to ensure that the voice actions that are performed while hands free cannot compromise the device's security. The activity should check the value of the EXTRA_SECURE extra to determine whether the device has been securely locked. If so, the activity should either restrict the set of voice actions that are permitted or require some form of secure authentication before proceeding. To ensure that the activity's user interface is visible while the lock screen is showing, the activity should set the FLAG_SHOW_WHEN_LOCKED window flag. Otherwise the activity's user interface may be hidden by the lock screen. The activity should take care not to leak private information when the device is securely locked.

Optional extras:

• EXTRA_SECURE

Starts an activity that will prompt the user for speech, send it through a speech recognizer, and either display a web search result or trigger another type of action based on the user's speech.

If you want to avoid triggering any type of action besides web search, you can use the EXTRA_WEB_SEARCH_ONLY extra.

Required extras:

• EXTRA_LANGUAGE_MODEL

Optional extras:

• EXTRA_PROMPT
• EXTRA_LANGUAGE
• EXTRA_MAX_RESULTS
• EXTRA_PARTIAL_RESULTS
• EXTRA_WEB_SEARCH_ONLY
• EXTRA_ORIGIN

Result extras (returned in the result, not to be specified in the request):

• EXTRA_RESULTS
• EXTRA_CONFIDENCE_SCORES (optional)

NOTE: There may not be any applications installed to handle this action, so you should make sure to catch ActivityNotFoundException.

Meta-data name under which an Activity implementing ACTION_WEB_SEARCH can use to expose the class name of a BroadcastReceiver which can respond to request for more information, from any of the broadcast intents specified in this class.

Broadcast intents can be directed to the class name specified in the meta-data by creating an Intent, setting the component with setComponent(android.content.ComponentName), and using sendOrderedBroadcast(Intent, String, BroadcastReceiver, android.os.Handler, int, String, android.os.Bundle) with another BroadcastReceiver which can receive the results.

The getVoiceDetailsIntent(Context) method is provided as a convenience to create a broadcast intent based on the value of this meta-data, if available.

This is optional and not all Activitys which implement ACTION_WEB_SEARCH are required to implement this. Thus retrieving this meta-data may be null.

The extra key used in an intent to the speech recognizer for voice search. Not generally to be used by developers. The system search dialog uses this, for example, to set a calling package for identification by a voice search API. If this extra is set by anyone but the system process, it should be overridden by the voice search implementation.

A float array of confidence scores of the recognition results when performing ACTION_RECOGNIZE_SPEECH. The array should be the same size as the ArrayList returned in EXTRA_RESULTS, and should contain values ranging from 0.0 to 1.0, or -1 to represent an unavailable confidence score.

Confidence values close to 1.0 indicate high confidence (the speech recognizer is confident that the recognition result is correct), while values close to 0.0 indicate low confidence.

Returned in the results; not to be specified in the recognition request. This extra is optional and might not be provided. Only present when RESULT_OK is returned in an activity result.

Optional IETF language tag (as defined by BCP 47), for example "en-US". This tag informs the recognizer to perform speech recognition in a language different than the one set in the getDefault().

Informs the recognizer which speech model to prefer when performing ACTION_RECOGNIZE_SPEECH. The recognizer uses this information to fine tune the results. This extra is required. Activities implementing ACTION_RECOGNIZE_SPEECH may interpret the values as they see fit.

The key to the extra in the Bundle returned by ACTION_GET_LANGUAGE_DETAILS which is a String that represents the current language preference this user has specified - a locale string like "en-US".

Optional limit on the maximum number of results to return. If omitted the recognizer will choose how many results to return. Must be an integer.

Specify this boolean extra in a broadcast of ACTION_GET_LANGUAGE_DETAILS to indicate that only the current language preference is needed in the response. This avoids any additional computation if all you need is EXTRA_LANGUAGE_PREFERENCE in the response.

Optional value which can be used to indicate the referer url of a page in which speech was requested. For example, a web browser may choose to provide this for uses of speech on a given page.

Optional boolean to indicate whether partial results should be returned by the recognizer as the user speaks (default is false). The server may ignore a request for partial results in some or all cases.

Optional text prompt to show to the user when asking them to speak.

An ArrayList<String> of the recognition results when performing ACTION_RECOGNIZE_SPEECH. Generally this list should be ordered in descending order of speech recognizer confidence. (See EXTRA_CONFIDENCE_SCORES). Returned in the results; not to be specified in the recognition request. Only present when RESULT_OK is returned in an activity result. In a PendingIntent, the lack of this extra indicates failure.

When the intent is ACTION_RECOGNIZE_SPEECH, the speech input activity will return results to you via the activity results mechanism. Alternatively, if you use this extra to supply a PendingIntent, the results will be added to its bundle and the PendingIntent will be sent to its target.

If you use EXTRA_RESULTS_PENDINGINTENT to supply a forwarding intent, you can also use this extra to supply additional extras for the final intent. The search results will be added to this bundle, and the combined bundle will be sent to the target.

Optional boolean to indicate that a "hands free" voice search was performed while the device was in a secure mode. An example of secure mode is when the device's screen lock is active, and it requires some form of authentication to be unlocked. When the device is securely locked, the voice search activity should either restrict the set of voice actions that are permitted, or require some form of secure authentication before proceeding.

The amount of time that it should take after we stop hearing speech to consider the input complete. Note that it is extremely rare you'd want to specify this value in an intent. If you don't have a very good reason to change these, you should leave them as they are. Note also that certain values may cause undesired or unexpected results - use judiciously! Additionally, depending on the recognizer implementation, these values may have no effect.

The minimum length of an utterance. We will not stop recording before this amount of time. Note that it is extremely rare you'd want to specify this value in an intent. If you don't have a very good reason to change these, you should leave them as they are. Note also that certain values may cause undesired or unexpected results - use judiciously! Additionally, depending on the recognizer implementation, these values may have no effect.

The amount of time that it should take after we stop hearing speech to consider the input possibly complete. This is used to prevent the endpointer cutting off during very short mid-speech pauses. Note that it is extremely rare you'd want to specify this value in an intent. If you don't have a very good reason to change these, you should leave them as they are. Note also that certain values may cause undesired or unexpected results - use judiciously! Additionally, depending on the recognizer implementation, these values may have no effect.

The key to the extra in the Bundle returned by ACTION_GET_LANGUAGE_DETAILS which is an ArrayList of Strings that represents the languages supported by this implementation of voice recognition - a list of strings like "en-US", "cmn-Hans-CN", etc.

Optional boolean, to be used with ACTION_WEB_SEARCH, to indicate whether to only fire web searches in response to a user's speech. The default is false, meaning that other types of actions can be taken based on the user's speech.

Use a language model based on free-form speech recognition. This is a value to use for EXTRA_LANGUAGE_MODEL.

Use a language model based on web search terms. This is a value to use for EXTRA_LANGUAGE_MODEL.

Result code returned when an audio error was encountered

Result code returned when there is a generic client error

Result code returned when a network error was encountered

Result code returned when no matches are found for the given speech

Result code returned when the recognition server returns an error

Returns the broadcast intent to fire with sendOrderedBroadcast(Intent, String, BroadcastReceiver, android.os.Handler, int, String, Bundle) to receive details from the package that implements voice search.

This is based on the value specified by the voice search Activity in DETAILS_META_DATA, and if this is not specified, will return null. Also if there is no chosen default to resolve for ACTION_WEB_SEARCH, this will return null.

If an intent is returned and is fired, a Bundle of extras will be returned to the provided result receiver, and should ideally contain values for EXTRA_LANGUAGE_PREFERENCE and EXTRA_SUPPORTED_LANGUAGES.

(Whether these are actually provided is up to the particular implementation. It is recommended that Activitys implementing ACTION_WEB_SEARCH provide this information, but it is not required.)