Request Syntax URI Request Parameters Request Body Response Syntax Response Elements Errors See Also

StartStreamTranscription

Starts a bidirectional HTTP/2 or WebSocket stream where audio is streamed to Amazon Transcribe and the transcription results are streamed to your application.

The following are encoded as headers:

language-code
media-encoding
sample-rate
session-id

For more information on streaming with Amazon Transcribe, see Transcribing streaming audio.

Request Syntax


POST /stream-transcription HTTP/2
x-amzn-transcribe-language-code: LanguageCode
x-amzn-transcribe-sample-rate: MediaSampleRateHertz
x-amzn-transcribe-media-encoding: MediaEncoding
x-amzn-transcribe-vocabulary-name: VocabularyName
x-amzn-transcribe-session-id: SessionId
x-amzn-transcribe-vocabulary-filter-name: VocabularyFilterName
x-amzn-transcribe-vocabulary-filter-method: VocabularyFilterMethod
x-amzn-transcribe-show-speaker-label: ShowSpeakerLabel
x-amzn-transcribe-enable-channel-identification: EnableChannelIdentification
x-amzn-transcribe-number-of-channels: NumberOfChannels
x-amzn-transcribe-enable-partial-results-stabilization: EnablePartialResultsStabilization
x-amzn-transcribe-partial-results-stability: PartialResultsStability
x-amzn-transcribe-content-identification-type: ContentIdentificationType
x-amzn-transcribe-content-redaction-type: ContentRedactionType
x-amzn-transcribe-pii-entity-types: PiiEntityTypes
x-amzn-transcribe-language-model-name: LanguageModelName
x-amzn-transcribe-identify-language: IdentifyLanguage
x-amzn-transcribe-language-options: LanguageOptions
x-amzn-transcribe-preferred-language: PreferredLanguage
x-amzn-transcribe-vocabulary-names: VocabularyNames
x-amzn-transcribe-vocabulary-filter-names: VocabularyFilterNames
Content-type: application/json

{
   "AudioStream": { 
      "AudioEvent": { 
         "AudioChunk": blob
      }
   }
}

URI Request Parameters

The request uses the following URI parameters.

ContentIdentificationType

Labels all personally identifiable information (PII) identified in your transcript.

Content identification is performed at the segment level; PII specified in PiiEntityTypes is flagged upon complete transcription of an audio segment.

You can’t set ContentIdentificationType and ContentRedactionType in the same request. If you set both, your request returns a BadRequestException.

For more information, see Redacting or identifying personally identifiable information.

Valid Values: PII

ContentRedactionType

Redacts all personally identifiable information (PII) identified in your transcript.

Content redaction is performed at the segment level; PII specified in PiiEntityTypes is redacted upon complete transcription of an audio segment.

You can’t set ContentRedactionType and ContentIdentificationType in the same request. If you set both, your request returns a BadRequestException.

For more information, see Redacting or identifying personally identifiable information.

Valid Values: PII

EnableChannelIdentification

Enables channel identification in multi-channel audio.

Channel identification transcribes the audio on each channel independently, then appends the output for each channel into one transcript.

If you have multi-channel audio and do not enable channel identification, your audio is transcribed in a continuous manner and your transcript is not separated by channel.

For more information, see Transcribing multi-channel audio.

EnablePartialResultsStabilization

Enables partial result stabilization for your transcription. Partial result stabilization can reduce latency in your output, but may impact accuracy. For more information, see Partial-result stabilization.

IdentifyLanguage

Enables automatic language identification for your transcription.

If you include IdentifyLanguage, you can optionally include a list of language codes, using LanguageOptions, that you think may be present in your audio stream. Including language options can improve transcription accuracy.

You can also include a preferred language using PreferredLanguage. Adding a preferred language can help Amazon Transcribe identify the language faster than if you omit this parameter.

If you have multi-channel audio that contains different languages on each channel, and you've enabled channel identification, automatic language identification identifies the dominant language on each audio channel.

Note that you must include either LanguageCode or IdentifyLanguage in your request. If you include both parameters, your request fails.

Streaming language identification can't be combined with custom language models or redaction.

LanguageCode

Specify the language code that represents the language spoken in your audio.

If you're unsure of the language spoken in your audio, consider using IdentifyLanguage to enable automatic language identification.

For a list of languages supported with Amazon Transcribe streaming, refer to the Supported languages table.

LanguageModelName

Specify the name of the custom language model that you want to use when processing your transcription. Note that language model names are case sensitive.

The language of the specified language model must match the language code you specify in your transcription request. If the languages don't match, the language model isn't applied. There are no errors or warnings associated with a language mismatch.

For more information, see Custom language models.

Length Constraints: Minimum length of 1. Maximum length of 200.

Pattern: ^[0-9a-zA-Z._-]+

LanguageOptions

Specify two or more language codes that represent the languages you think may be present in your media; including more than five is not recommended. If you're unsure what languages are present, do not include this parameter.

Including language options can improve the accuracy of language identification.

If you include LanguageOptions in your request, you must also include IdentifyLanguage.

For a list of languages supported with Amazon Transcribe streaming, refer to the Supported languages table.

Important

You can only include one language dialect per language per stream. For example, you cannot include en-US and en-AU in the same request.

Length Constraints: Minimum length of 1. Maximum length of 200.

Pattern: ^[a-zA-Z-,]+

MediaEncoding

Specify the encoding used for the input audio. Supported formats are:

FLAC
OPUS-encoded audio in an Ogg container
PCM (only signed 16-bit little-endian audio formats, which does not include WAV)

For more information, see Media formats.

Valid Values: pcm | ogg-opus | flac

Required: Yes

MediaSampleRateHertz

The sample rate of the input audio (in hertz). Low-quality audio, such as telephone audio, is typically around 8,000 Hz. High-quality audio typically ranges from 16,000 Hz to 48,000 Hz. Note that the sample rate you specify must match that of your audio.

Valid Range: Minimum value of 8000. Maximum value of 48000.

Required: Yes

NumberOfChannels

Specify the number of channels in your audio stream. Up to two channels are supported.

Valid Range: Minimum value of 2.

PartialResultsStability

Specify the level of stability to use when you enable partial results stabilization (EnablePartialResultsStabilization).

Low stability provides the highest accuracy. High stability transcribes faster, but with slightly lower accuracy.

For more information, see Partial-result stabilization.

Valid Values: high | medium | low

PiiEntityTypes

Specify which types of personally identifiable information (PII) you want to redact in your transcript. You can include as many types as you'd like, or you can select ALL.

To include PiiEntityTypes in your request, you must also include either ContentIdentificationType or ContentRedactionType.

Values must be comma-separated and can include: BANK_ACCOUNT_NUMBER, BANK_ROUTING, CREDIT_DEBIT_NUMBER, CREDIT_DEBIT_CVV, CREDIT_DEBIT_EXPIRY, PIN, EMAIL, ADDRESS, NAME, PHONE, SSN, or ALL.

Length Constraints: Minimum length of 1. Maximum length of 300.

Pattern: ^[A-Z_, ]+

PreferredLanguage

Specify a preferred language from the subset of languages codes you specified in LanguageOptions.

You can only use this parameter if you've included IdentifyLanguage and LanguageOptions in your request.

SessionId

Specify a name for your transcription session. If you don't include this parameter in your request, Amazon Transcribe generates an ID and returns it in the response.

You can use a session ID to retry a streaming session.

Length Constraints: Fixed length of 36.

Pattern: [a-fA-F0-9]{8}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{12}

ShowSpeakerLabel

Enables speaker partitioning (diarization) in your transcription output. Speaker partitioning labels the speech from individual speakers in your media file.

For more information, see Partitioning speakers (diarization).

VocabularyFilterMethod

Specify how you want your vocabulary filter applied to your transcript.

To replace words with ***, choose mask.

To delete words, choose remove.

To flag words without changing them, choose tag.

Valid Values: remove | mask | tag

VocabularyFilterName

Specify the name of the custom vocabulary filter that you want to use when processing your transcription. Note that vocabulary filter names are case sensitive.

If the language of the specified custom vocabulary filter doesn't match the language identified in your media, your job fails.

Important

This parameter is not intended for use with the IdentifyLanguage parameter. If you're including IdentifyLanguage in your request and want to use one or more vocabulary filters with your transcription, use the VocabularyFilterNames parameter instead.

For more information, see Using vocabulary filtering with unwanted words.

Length Constraints: Minimum length of 1. Maximum length of 200.

Pattern: ^[0-9a-zA-Z._-]+

VocabularyFilterNames

Specify the names of the custom vocabulary filters that you want to use when processing your transcription. Note that vocabulary filter names are case sensitive.

If none of the languages of the specified custom vocabulary filters match the language identified in your media, your job fails.

Important

This parameter is only intended for use with the IdentifyLanguage parameter. If you're not including IdentifyLanguage in your request and want to use a custom vocabulary filter with your transcription, use the VocabularyFilterName parameter instead.

For more information, see Using vocabulary filtering with unwanted words.

Length Constraints: Minimum length of 1. Maximum length of 3000.

Pattern: ^[a-zA-Z0-9,-._]+

VocabularyName

Specify the name of the custom vocabulary that you want to use when processing your transcription. Note that vocabulary names are case sensitive.

If the language of the specified custom vocabulary doesn't match the language identified in your media, your job fails.

Important

This parameter is not intended for use with the IdentifyLanguage parameter. If you're including IdentifyLanguage in your request and want to use one or more custom vocabularies with your transcription, use the VocabularyNames parameter instead.

For more information, see Custom vocabularies.

Length Constraints: Minimum length of 1. Maximum length of 200.

Pattern: ^[0-9a-zA-Z._-]+

VocabularyNames

Specify the names of the custom vocabularies that you want to use when processing your transcription. Note that vocabulary names are case sensitive.

If none of the languages of the specified custom vocabularies match the language identified in your media, your job fails.

Important

This parameter is only intended for use with the IdentifyLanguage parameter. If you're not including IdentifyLanguage in your request and want to use a custom vocabulary with your transcription, use the VocabularyName parameter instead.

For more information, see Custom vocabularies.

Length Constraints: Minimum length of 1. Maximum length of 3000.

Pattern: ^[a-zA-Z0-9,-._]+

Request Body

The request accepts the following data in JSON format.

AudioStream

An encoded stream of audio blobs. Audio streams are encoded as either HTTP/2 or WebSocket data frames.

For more information, see Transcribing streaming audio.

Type: AudioStream object

Required: Yes

Response Syntax


HTTP/2 200
x-amzn-request-id: RequestId
x-amzn-transcribe-language-code: LanguageCode
x-amzn-transcribe-sample-rate: MediaSampleRateHertz
x-amzn-transcribe-media-encoding: MediaEncoding
x-amzn-transcribe-vocabulary-name: VocabularyName
x-amzn-transcribe-session-id: SessionId
x-amzn-transcribe-vocabulary-filter-name: VocabularyFilterName
x-amzn-transcribe-vocabulary-filter-method: VocabularyFilterMethod
x-amzn-transcribe-show-speaker-label: ShowSpeakerLabel
x-amzn-transcribe-enable-channel-identification: EnableChannelIdentification
x-amzn-transcribe-number-of-channels: NumberOfChannels
x-amzn-transcribe-enable-partial-results-stabilization: EnablePartialResultsStabilization
x-amzn-transcribe-partial-results-stability: PartialResultsStability
x-amzn-transcribe-content-identification-type: ContentIdentificationType
x-amzn-transcribe-content-redaction-type: ContentRedactionType
x-amzn-transcribe-pii-entity-types: PiiEntityTypes
x-amzn-transcribe-language-model-name: LanguageModelName
x-amzn-transcribe-identify-language: IdentifyLanguage
x-amzn-transcribe-language-options: LanguageOptions
x-amzn-transcribe-preferred-language: PreferredLanguage
x-amzn-transcribe-vocabulary-names: VocabularyNames
x-amzn-transcribe-vocabulary-filter-names: VocabularyFilterNames
Content-type: application/json

{
   "TranscriptResultStream": { 
      "BadRequestException": { 
      },
      "ConflictException": { 
      },
      "InternalFailureException": { 
      },
      "LimitExceededException": { 
      },
      "ServiceUnavailableException": { 
      },
      "TranscriptEvent": { 
         "Transcript": { 
            "Results": [ 
               { 
                  "Alternatives": [ 
                     { 
                        "Entities": [ 
                           { 
                              "Category": "string",
                              "Confidence": number,
                              "Content": "string",
                              "EndTime": number,
                              "StartTime": number,
                              "Type": "string"
                           }
                        ],
                        "Items": [ 
                           { 
                              "Confidence": number,
                              "Content": "string",
                              "EndTime": number,
                              "Speaker": "string",
                              "Stable": boolean,
                              "StartTime": number,
                              "Type": "string",
                              "VocabularyFilterMatch": boolean
                           }
                        ],
                        "Transcript": "string"
                     }
                  ],
                  "ChannelId": "string",
                  "EndTime": number,
                  "IsPartial": boolean,
                  "LanguageCode": "string",
                  "LanguageIdentification": [ 
                     { 
                        "LanguageCode": "string",
                        "Score": number
                     }
                  ],
                  "ResultId": "string",
                  "StartTime": number
               }
            ]
         }
      }
   }
}

Response Elements

If the action is successful, the service sends back an HTTP 200 response.

The response returns the following HTTP headers.

ContentIdentificationType

Shows whether content identification was enabled for your transcription.

Valid Values: PII

ContentRedactionType

Shows whether content redaction was enabled for your transcription.

Valid Values: PII

EnableChannelIdentification

Shows whether channel identification was enabled for your transcription.

EnablePartialResultsStabilization

Shows whether partial results stabilization was enabled for your transcription.

IdentifyLanguage

Shows whether automatic language identification was enabled for your transcription.

LanguageCode

Provides the language code that you specified in your request.

LanguageModelName

Provides the name of the custom language model that you specified in your request.

Length Constraints: Minimum length of 1. Maximum length of 200.

Pattern: ^[0-9a-zA-Z._-]+

LanguageOptions

Provides the language codes that you specified in your request.

Length Constraints: Minimum length of 1. Maximum length of 200.

Pattern: ^[a-zA-Z-,]+

MediaEncoding

Provides the media encoding you specified in your request.

Valid Values: pcm | ogg-opus | flac

MediaSampleRateHertz

Provides the sample rate that you specified in your request.

Valid Range: Minimum value of 8000. Maximum value of 48000.

NumberOfChannels

Provides the number of channels that you specified in your request.

Valid Range: Minimum value of 2.

PartialResultsStability

Provides the stabilization level used for your transcription.

Valid Values: high | medium | low

PiiEntityTypes

Lists the PII entity types you specified in your request.

Length Constraints: Minimum length of 1. Maximum length of 300.

Pattern: ^[A-Z_, ]+

PreferredLanguage

Provides the preferred language that you specified in your request.

RequestId

Provides the identifier for your streaming request.

SessionId

Provides the identifier for your transcription session.

Length Constraints: Fixed length of 36.

Pattern: [a-fA-F0-9]{8}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{4}-[a-fA-F0-9]{12}

ShowSpeakerLabel

Shows whether speaker partitioning was enabled for your transcription.

VocabularyFilterMethod

Provides the vocabulary filtering method used in your transcription.

Valid Values: remove | mask | tag

VocabularyFilterName

Provides the name of the custom vocabulary filter that you specified in your request.

Length Constraints: Minimum length of 1. Maximum length of 200.

Pattern: ^[0-9a-zA-Z._-]+

VocabularyFilterNames

Provides the names of the custom vocabulary filters that you specified in your request.

Length Constraints: Minimum length of 1. Maximum length of 3000.

Pattern: ^[a-zA-Z0-9,-._]+

VocabularyName

Provides the name of the custom vocabulary that you specified in your request.

Length Constraints: Minimum length of 1. Maximum length of 200.

Pattern: ^[0-9a-zA-Z._-]+

VocabularyNames

Provides the names of the custom vocabularies that you specified in your request.

Length Constraints: Minimum length of 1. Maximum length of 3000.

Pattern: ^[a-zA-Z0-9,-._]+

The following data is returned in JSON format by the service.

TranscriptResultStream

Provides detailed information about your streaming session.

Type: TranscriptResultStream object

Errors

For information about the errors that are common to all actions, see Common Errors.

BadRequestException

One or more arguments to the StartStreamTranscription or StartMedicalStreamTranscription operation was not valid. For example, MediaEncoding or LanguageCode used not valid values. Check the specified parameters and try your request again.

HTTP Status Code: 400

ConflictException

A new stream started with the same session ID. The current stream has been terminated.

HTTP Status Code: 409

InternalFailureException

A problem occurred while processing the audio. Amazon Transcribe terminated processing.

HTTP Status Code: 500

LimitExceededException

Your client has exceeded one of the Amazon Transcribe limits. This is typically the audio length limit. Break your audio stream into smaller chunks and try your request again.

HTTP Status Code: 429

ServiceUnavailableException

The service is currently unavailable. Try your request later.

HTTP Status Code: 503