gRPC & Protocol Buffers

riva/proto/riva_asr.proto

service RivaSpeechRecognition

The RivaSpeechRecognition service provides two mechanisms for converting speech to text.

rpc RecognizeResponse Recognize(RecognizeRequest)

Recognize expects a RecognizeRequest and returns a RecognizeResponse. This request will block until the audio is uploaded, processed, and a transcript is returned.

rpc stream StreamingRecognizeResponse StreamingRecognize(stream StreamingRecognizeRequest)

StreamingRecognize is a non-blocking API call that allows audio data to be fed to the server in chunks as it becomes available. Depending on the configuration in the StreamingRecognizeRequest, intermediate results can be sent back to the client. Recognition ends when the stream is closed by the client.

message RecognitionConfig

Provides information to the recognizer that specifies how to process the request

nvidia.riva.AudioEncoding encoding

The encoding of the audio data sent in the request.

All encodings support only 1 channel (mono) audio.

int32 sample_rate_hertz

Sample rate in Hertz of the audio data sent in all RecognizeAudio messages.

string language_code

Required. The language of the supplied audio as a [BCP-47](https://www.rfc-editor.org/rfc/bcp/bcp47.txt) language tag. Example: “en-US”. Currently only en-US is supported

int32 max_alternatives

Maximum number of recognition hypotheses to be returned. Specifically, the maximum number of SpeechRecognizeAlternative messages within each SpeechRecognizeResult. The server may return fewer than max_alternatives. If omitted, will return a maximum of one.

SpeechContext speech_contexts(repeated)

Array of SpeechContext. A means to provide context to assist the speech recognition. For more information, see SpeechContext section

int32 audio_channel_count

The number of channels in the input audio data. ONLY set this for MULTI-CHANNEL recognition. Valid values for LINEAR16 and FLAC are 1-8. Valid values for OGG_OPUS are ‘1’-‘254’. Valid value for MULAW, AMR, AMR_WB and SPEEX_WITH_HEADER_BYTE is only 1. If 0 or omitted, defaults to one channel (mono). Note: We only recognize the first channel by default. To perform independent recognition on each channel set enable_separate_recognition_per_channel to ‘true’.

bool enable_word_time_offsets

If true, the top result includes a list of words and the start and end time offsets (timestamps) for those words. If false, no word-level time offset information is returned. The default is false.

bool enable_automatic_punctuation

If ‘true’, adds punctuation to recognition result hypotheses. The default ‘false’ value does not add punctuation to result hypotheses.

bool enable_separate_recognition_per_channel

This needs to be set to true explicitly and audio_channel_count > 1 to get each channel recognized separately. The recognition result will contain a channel_tag field to state which channel that result belongs to. If this is not true, we will only recognize the first channel. The request is billed cumulatively for all channels recognized: audio_channel_count multiplied by the length of the audio.

string model

Which model to select for the given request. Valid choices: Jasper, Quartznet

bool verbatim_transcripts

The verbatim_transcripts flag enables or disable inverse text normalization. ‘true’ returns exactly what was said, with no denormalization. ‘false’ applies inverse text normalization, also this is the default

RecognitionConfig.CustomConfigurationEntry custom_configuration (repeated)

Custom fields for passing request-level configuration options to plugins used in the model pipeline.

message RecognitionConfig.CustomConfigurationEntry

string key

string value

message RecognizeRequest

RecognizeRequest is used for batch processing of a single audio recording.

RecognitionConfig config

Provides information to recognizer that specifies how to process the request.

bytes audio

The raw audio data to be processed. The audio bytes must be encoded as specified in RecognitionConfig.

message RecognizeResponse

The only message returned to the client by the Recognize method. It contains the result as zero or more sequential SpeechRecognitionResult messages.

SpeechRecognitionResult results(repeated)

Sequential list of transcription results corresponding to sequential portions of audio. Currently only returns one transcript.

message SpeechContext

Provides “hints” to the speech recognizer to favor specific words and phrases in the results.

string phrases(repeated)

A list of strings containing words so that the speech recognition is more likely to recognize them. This can be used to improve the accuracy for specific words, for example, if specific commands are typically spoken by the user.

float boost

Hint Boost. Positive value will increase the probability that a specific phrase will be recognized over other similar sounding phrases. The higher the boost, the higher the chance of false positive recognition as well. Though boost can accept a wide range of positive values, most use cases are best served with values between 0 and 20. We recommend using a binary search approach to finding the optimal value for your use case.

message SpeechRecognitionAlternative

Alternative hypotheses (a.k.a. n-best list).

string transcript

Transcript text representing the words that the user spoke.

float confidence

The non-normalized confidence estimate. A higher number indicates an estimated greater likelihood that the recognized words are correct. This field is set only for a non-streaming result or, of a streaming result where is_final=true. This field is not guaranteed to be accurate and users should not rely on it to be always provided.

WordInfo words(repeated)

A list of word-specific information for each recognized word. Only populated if is_final=true

message SpeechRecognitionResult

A speech recognition result corresponding to the latest transcript

SpeechRecognitionAlternative alternatives(repeated)

May contain one or more recognition hypotheses (up to the maximum specified in max_alternatives). These alternatives are ordered in terms of accuracy, with the top (first) alternative being the most probable, as ranked by the recognizer.

int32 channel_tag

For multi-channel audio, this is the channel number corresponding to the recognized result for the audio from that channel. For audio_channel_count = N, its output values can range from ‘1’ to ‘N’.

float audio_processed

Length of audio processed so far in seconds

message StreamingRecognitionConfig

Provides information to the recognizer that specifies how to process the request

RecognitionConfig config

Provides information to the recognizer that specifies how to process the request

bool interim_results

If true, interim results (tentative hypotheses) may be returned as they become available (these interim results are indicated with the is_final=false flag). If false or omitted, only is_final=true result(s) are returned.

message StreamingRecognitionResult

A streaming speech recognition result corresponding to a portion of the audio that is currently being processed.

SpeechRecognitionAlternative alternatives(repeated)

May contain one or more recognition hypotheses (up to the maximum specified in max_alternatives). These alternatives are ordered in terms of accuracy, with the top (first) alternative being the most probable, as ranked by the recognizer.

bool is_final

If false, this StreamingRecognitionResult represents an interim result that may change. If true, this is the final time the speech service will return this particular StreamingRecognitionResult, the recognizer will not return any further hypotheses for this portion of the transcript and corresponding audio.

float stability

An estimate of the likelihood that the recognizer will not change its guess about this interim result. Values range from 0.0 (completely unstable) to 1.0 (completely stable). This field is only provided for interim results (is_final=false). The default of 0.0 is a sentinel value indicating stability was not set.

int32 channel_tag

For multi-channel audio, this is the channel number corresponding to the recognized result for the audio from that channel. For audio_channel_count = N, its output values can range from ‘1’ to ‘N’.

float audio_processed

Length of audio processed so far in seconds

message StreamingRecognizeRequest

A StreamingRecognizeRequest is used to configure and stream audio content to the Riva ASR Service. The first message sent must include only a StreamingRecognitionConfig. Subsequent messages sent in the stream must contain only raw bytes of the audio to be recognized.

StreamingRecognitionConfig streaming_config

Provides information to the recognizer that specifies how to process the request. The first StreamingRecognizeRequest message must contain a streaming_config message.

bytes audio_content

The audio data to be recognized. Sequential chunks of audio data are sent in sequential StreamingRecognizeRequest messages. The first StreamingRecognizeRequest message must not contain audio data and all subsequent StreamingRecognizeRequest messages must contain audio data. The audio bytes must be encoded as specified in RecognitionConfig.

message StreamingRecognizeResponse

StreamingRecognitionResult results(repeated)

This repeated list contains the latest transcript(s) corresponding to audio currently being processed. Currently one result is returned, where each result can have multiple alternatives

message WordInfo

Word-specific information for recognized words.

int32 start_time

Time offset relative to the beginning of the audio in ms and corresponding to the start of the spoken word. This field is only set if enable_word_time_offsets=true and only in the top hypothesis.

int32 end_time

Time offset relative to the beginning of the audio in ms and corresponding to the end of the spoken word. This field is only set if enable_word_time_offsets=true and only in the top hypothesis.

string word

The word corresponding to this set of information.

riva/proto/riva_nlp.proto

service RivaLanguageUnderstanding

rpc TextClassResponse ClassifyText(TextClassRequest)

ClassifyText takes as input an input/query string and parameters related to the requested model to use to evaluate the text. The service evaluates the text with the requested model, and returns one or more classifications.

rpc TokenClassResponse ClassifyTokens(TokenClassRequest)

ClassifyTokens takes as input either a string or list of tokens and parameters related to which model to use. The service evaluates the text with the requested model, performing additional tokenization if necessary, and returns one or more class labels per token.

rpc TextTransformResponse TransformText(TextTransformRequest)

TransformText takes an input/query string and parameters related to the requested model and returns another string. The behavior of the function is defined entirely by the underlying model and may be used for tasks like translation, adding punctuation, augment the input directly, etc.

rpc TokenClassResponse AnalyzeEntities(AnalyzeEntitiesRequest)

AnalyzeEntities accepts an input string and returns all named entities within the text, as well as a category and likelihood.

rpc AnalyzeIntentResponse AnalyzeIntent(AnalyzeIntentRequest)

AnalyzeIntent accepts an input string and returns the most likely intent as well as slots relevant to that intent.

The model requires that a valid “domain” be passed in, and optionally supports including a previous intent classification result to provide context for the model.

rpc TextTransformResponse PunctuateText(TextTransformRequest)

PunctuateText takes text with no- or limited- punctuation and returns the same text with corrected punctuation and capitalization.

rpc NaturalQueryResponse NaturalQuery(NaturalQueryRequest)

NaturalQuery is a search function that enables querying one or more documents or contexts with a query that is written in natural language.

message AnalyzeEntitiesOptions

AnalyzeEntitiesOptions is an optional configuration message to be sent as part of an AnalyzeEntitiesRequest with query metadata

string lang

Optional language field. Assumed to be “en-US” if not specified.

message AnalyzeEntitiesRequest

AnalyzeEntitiesRequest is the input message for the AnalyzeEntities service

string query

The string to analyze for intent and slots

AnalyzeEntitiesOptions options

Optional configuration for the request, including providing context from previous turns and hardcoding a domain/language

message AnalyzeIntentContext

AnalyzeIntentContext is reserved for future use when we may send context back in a a variety of different formats (including raw neural network hidden states)

Reserved for future use

message AnalyzeIntentOptions

AnalyzeIntentOptions is an optional configuration message to be sent as part of an AnalyzeIntentRequest with query metadata

string previous_intent

AnalyzeIntentContext vectors

string domain

Optional domain field. Domain must be supported otherwise an error will be returned. If left blank, a domain detector will be run first and then the query routed to the appropriate intent classifier (if it exists)

string lang

Optional language field. Assumed to be “en-US” if not specified.

message AnalyzeIntentRequest

AnalyzeIntentRequest is the input message for the AnalyzeIntent service

string query

The string to analyze for intent and slots

AnalyzeIntentOptions options

Optional configuration for the request, including providing context from previous turns and hardcoding a domain/language

message AnalyzeIntentResponse

AnalyzeIntentResponse is returned by the AnalyzeIntent service, and includes information related to the query’s intent, (optionally) slot data, and its domain.

Classification intent

Intent classification result, including the label and score

TokenClassValue slots(repeated)

List of tokens explicitly marked as filling a slot relevant to the intent, where the tokens may not exactly match the input (based on the recombined values after tokenization)

string domain_str

Returns the inferred domain for the query if not hardcoded in the request. In the case where the domain was hardcoded in AnalyzeIntentRequest, the returned domain is an exact match to the request. In the case where no domain matches the query, intent and slots will be unset.

DEPRECATED, use Classification domain field.

Classification domain

Returns the inferred domain for the query if not hardcoded in the request. In the case where the domain was hardcoded in AnalyzeIntentRequest, the returned domain is an exact match to the request. In the case where no domain matches the query, intent and slots will be unset.

message Classification

Classification messages return a class name and corresponding score

string class_name

float score

message ClassificationResult

ClassificationResults contain zero or more Classification messages If the number of Classifications is > 1, top_n > 1 must have been specified.

Classification labels(repeated)

message NLPModelParams

NLPModelParams is a metadata message that is included in every request message used by the Core NLP Service and is used to specify model characteristics/requirements

string model_name

Requested model to use. If unavailable, the request will return an error

message NaturalQueryRequest

string query

The natural language query

uint32 top_n

Maximum number of answers to return for the query. Defaults to 1 if not set.

string context

Context to search with the above query

message NaturalQueryResponse

NaturalQueryResult results(repeated)

message NaturalQueryResult

string answer

text which answers the query

float score

Score representing confidence in result

message Span

Span of a particular result

uint32 start

uint32 end

message TextClassRequest

TextClassRequest is the input message to the ClassifyText service.

string text(repeated)

Each repeated text element is handled independently for handling multiple input strings with a single request

uint32 top_n

Return the top N classification results for each input. 0 or 1 will return top class, otherwise N. Note: Current disabled.

NLPModelParams model

message TextClassResponse

TextClassResponse is the return message from the ClassifyText service.

ClassificationResult results(repeated)

message TextTransformRequest

TextTransformRequest is a request type intended for services like TransformText which take an arbitrary text input

string text(repeated)

Each repeated text element is handled independently for handling multiple input strings with a single request

uint32 top_n

NLPModelParams model

message TextTransformResponse

TextTransformResponse is returned by the TransformText method. Responses are returned in the same order as they were requested.

string text(repeated)

message TokenClassRequest

TokenClassRequest is the input message to the ClassifyText service.

string text(repeated)

Each repeated text element is handled independently for handling multiple input strings with a single request

uint32 top_n

Return the top N classification results for each input. 0 or 1 will return top class, otherwise N. Note: Current disabled.

NLPModelParams model

message TokenClassResponse

TokenClassResponse returns a single TokenClassSequence per input request

TokenClassSequence results(repeated)

message TokenClassSequence

TokenClassSequence is used for returning a sequence of TokenClassValue objects in the original order of input tokens

TokenClassValue results(repeated)

message TokenClassValue

TokenClassValue is used to correlate an input token with its classification results

string token

Classification label(repeated)

Span span(repeated)

riva/proto/riva_tts.proto

service RivaSpeechSynthesis

rpc SynthesizeSpeechResponse Synthesize(SynthesizeSpeechRequest)

Used to request text-to-speech from the service. Submit a request containing the desired text and configuration, and receive audio bytes in the requested format.

rpc stream SynthesizeSpeechResponse SynthesizeOnline(SynthesizeSpeechRequest)

Used to request text-to-speech returned via stream as it becomes available. Submit a SynthesizeSpeechRequest with desired text and configuration, and receive stream of bytes in the requested format.

message SynthesizeSpeechRequest

string text

string language_code

nvidia.riva.AudioEncoding encoding

audio encoding params

int32 sample_rate_hz

string voice_name

voice params

message SynthesizeSpeechResponse

bytes audio

SynthesizeSpeechResponseMetadata meta

message SynthesizeSpeechResponseMetadata

string text

Currently experimental API addition that returns the input text after preprocessing has been completed as well as the predicted duration for each token. Note: this message is subject to future breaking changes, and potential removal.

string processed_text

float predicted_durations(repeated)

riva/proto/riva_audio.proto

enum AudioEncoding

AudioEncoding specifies the encoding of the audio bytes in the encapsulating message.

enumerator ENCODING_UNSPECIFIED = 0

Not specified.

enumerator LINEAR_PCM = 1

Uncompressed 16-bit signed little-endian samples (Linear PCM).

enumerator FLAC = 2

FLAC (Free Lossless Audio Codec) is the recommended encoding because it is lossless–therefore recognition is not compromised–and requires only about half the bandwidth of LINEAR16. FLAC stream encoding supports 16-bit and 24-bit samples, however, not all fields in STREAMINFO are supported.

enumerator MULAW = 3

8-bit samples that compand 14-bit audio samples using G.711 PCMU/mu-law.

enumerator ALAW = 20

8-bit samples that compand 13-bit audio samples using G.711 PCMU/a-law.

riva/proto/health.proto

service Health

rpc HealthCheckResponse Check(HealthCheckRequest)

rpc stream HealthCheckResponse Watch(HealthCheckRequest)

message HealthCheckRequest

string service

message HealthCheckResponse

HealthCheckResponse.ServingStatus status

enum HealthCheckResponse.ServingStatus

enumerator UNKNOWN = 0

enumerator SERVING = 1

enumerator NOT_SERVING = 2