@deepgram/sdk
Advanced tools
Official Node.js SDK for Deepgram's automated speech recognition APIs.
To access the API you will need a Deepgram account. Sign up for free at signup.
You can learn more about the full Deepgram API at https://developers.deepgram.com.
npm install @deepgram/sdk
yarn add @deepgram/sdk
const { Deepgram } = require("@deepgram/sdk");
const deepgram = new Deepgram({
apiKey: DEEPGRAM_API_KEY,
apiUrl: CUSTOM_API_URL, // Optionally used for on-premises customers
});
The transcription property can handle both pre-recorded and live transcriptions.
The transcription.preRecorded method handles sending an existing file or
buffer to the Deepgram API to generate a transcription. Additional options
can be provided to customize the result.
// Sending a file
const fileSource = { url: URL_OF_FILE };
// Sending a buffer
const bufferSource = { buffer: BUFFER_OF_FILE, mimetype: MIMETYPE_OF_FILE };
// Both fileSource or bufferSource could be provided as the source parameter
const response = await deepgram.transcription.preRecorded(
fileSource | bufferSource,
{
punctuate: true,
// other options are available
}
);
Additional transcription options can be provided for prerecorded transcriptions.
{
/**
* AI model used to process submitted audio.
* @default general
* @remarks Possible values are general, phonecall, meeting or a custom string
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/model
*/
model?: Models | string;
/**
* Version of the model to use.
* @default latest
* @remarks latest OR <version_id>
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/version
*/
version: string;
/**
* BCP-47 language tag that hints at the primary spoken language.
* @default en-US
* @remarks Possible values are en-GB, en-IN, en-NZ, en-US, es, fr, ko, pt,
* pt-BR, ru, tr or null
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/language
*/
language?: string;
/**
* Indicates whether to add punctuation and capitalization to the transcript.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/punctuate
*/
punctuate?: boolean;
/**
* Indicates whether to remove profanity from the transcript.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/profanity_filter
*/
profanity_filter?: boolean;
/**
* Indicates whether to redact sensitive information, replacing redacted content with asterisks (*).
* @remarks Options include:
* `pci`: Redacts sensitive credit card information, including credit card number, expiration date, and CVV
* `numbers` (or `true)`: Aggressively redacts strings of numerals
* `ssn` (*beta*): Redacts social security numbers
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/redact
*/
redact?: Array<string>;
/**
* Indicates whether to recognize speaker changes. When set to true, each word
* in the transcript will be assigned a speaker number starting at 0.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/diarize
*/
diarize?: boolean;
/**
* Indicates whether to transcribe each audio channel independently. When set
* to true, you will receive one transcript for each channel, which means you
* can apply a different model to each channel using the model parameter (e.g.,
* set model to general:phonecall, which applies the general model to channel
* 0 and the phonecall model to channel 1).
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/multichannel
*/
multichannel?: boolean;
/**
* Maximum number of transcript alternatives to return. Just like a human listener,
* Deepgram can provide multiple possible interpretations of what it hears.
* @default 1
*/
alternatives?: number;
/**
* Indicates whether to convert numbers from written format (e.g., one) to
* numerical format (e.g., 1). Deepgram can format numbers up to 999,999.
* @remarks Converted numbers do not include punctuation. For example,
* 999,999 would be transcribed as 999999.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/numerals
*/
numerals?: boolean;
/**
* Terms or phrases to search for in the submitted audio. Deepgram searches
* for acoustic patterns in audio rather than text patterns in transcripts
* because we have noticed that acoustic pattern matching is more performant.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/search
*/
search?: Array<string>;
/**
* Callback URL to provide if you would like your submitted audio to be
* processed asynchronously. When passed, Deepgram will immediately respond
* with a request_id. When it has finished analyzing the audio, it will send
* a POST request to the provided URL with an appropriate HTTP status code.
* @remarks You may embed basic authentication credentials in the callback URL.
* Only ports 80, 443, 8080, and 8443 can be used for callbacks.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/callback
*/
callback?: string;
/**
* Keywords to which the model should pay particular attention to boosting
* or suppressing to help it understand context. Just like a human listener,
* Deepgram can better understand mumbled, distorted, or otherwise
* hard-to-decipher speech when it knows the context of the conversation.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/keywords
*/
keywords?: Array<string>;
/**
* Indicates whether Deepgram will segment speech into meaningful semantic
* units, which allows the model to interact more naturally and effectively
* with speakers' spontaneous speech patterns. For example, when humans
* speak to each other conversationally, they often pause mid-sentence to
* reformulate their thoughts, or stop and restart a badly-worded sentence.
* When utterances is set to true, these utterances are identified and
* returned in the transcript results.
*
* By default, when utterances is enabled, it starts a new utterance after
* 0.8 s of silence. You can customize the length of time used to determine
* where to split utterances by submitting the utt_split parameter.
* @remarks **BETA FEATURE**
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/utterances
*/
utterances?: boolean;
/**
* Length of time in seconds of silence between words that Deepgram will
* use when determining where to split utterances. Used when utterances
* is enabled.
* @default 0.8 seconds
* @remarks **BETA FEATURE**
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/utt_split
*/
utt_split?: number;
}
{
request_id?: string;
metadata?: {
request_id: string;
transaction_key: string;
sha256: string;
created: string;
duration: number;
channels: number;
};
results?: {
channels: Array<{
search?: Array<{
query: string;
hits: Array<{
confidence: number;
start: number;
end: number;
snippet: string;
}>;
}>;
alternatives: Array<{
transcript: string;
confidence: number;
words: Array<{
word: string;
start: number;
end: number;
confidence: number;
punctuated_word?: string;
}>;
}>;
}>;
utterances?: Array<{
start: number;
end: number;
confidence: number;
channel: number;
transcript: string;
words: Array<{
word: string;
start: number;
end: number;
confidence: number;
punctuated_word?: string;
}>;
speaker?: number;
id: string;
}>;
};
};
The transcription.live method provides access to a websocket connection
to the Deepgram API for generating streaming transcriptions. Additional options
can be provided to customize the result.
const deepgramLive = deepgram.transcription.live({ punctuate: true });
socket.on("microphone-stream", (stream) => {
deepgramSocket.send(stream);
});
/**
* Receive transcriptions based on sent streams
*/
deepgramLive.addListener("transcriptReceived", (transcription) => {
console.log(transcription.data);
});
The following events are fired by the live transcription object:
| Event | Description | Data |
|---|---|---|
open | The websocket connection to Deepgram has been opened. | The DG live transcription object |
close | The websocket connection to Deepgram has been closed. | WebSocket.CloseEvent |
error | An error occurred with the websocket connection | Error object |
transcriptReceived | Deepgram has responded with a transcription | Transcription Response |
Additional transcription options can be provided for live transcriptions.
{
/**
* AI model used to process submitted audio.
* @default general
* @remarks Possible values are general, phonecall, meeting or a custom string
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/model
*/
model?: Models | string;
/**
* Version of the model to use.
* @default latest
* @remarks latest OR <version_id>
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/version
*/
version: string;
/**
* BCP-47 language tag that hints at the primary spoken language.
* @default en-US
* @remarks Possible values are en-GB, en-IN, en-NZ, en-US, es, fr, ko, pt,
* pt-BR, ru, tr or null
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/language
*/
language?: string;
/**
* Indicates whether to add punctuation and capitalization to the transcript.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/punctuate
*/
punctuate?: boolean;
/**
* Indicates whether to remove profanity from the transcript.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/profanity_filter
*/
profanity_filter?: boolean;
/**
* Indicates whether to redact sensitive information, replacing redacted content with asterisks (*).
* @remarks Options include:
* `pci`: Redacts sensitive credit card information, including credit card number, expiration date, and CVV
* `numbers` (or `true)`: Aggressively redacts strings of numerals
* `ssn` (*beta*): Redacts social security numbers
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/redact
*/
redact?: Array<string>;
/**
* Indicates whether to recognize speaker changes. When set to true, each word
* in the transcript will be assigned a speaker number starting at 0.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/diarize
*/
diarize?: boolean;
/**
* Indicates whether to transcribe each audio channel independently. When set
* to true, you will receive one transcript for each channel, which means you
* can apply a different model to each channel using the model parameter (e.g.,
* set model to general:phonecall, which applies the general model to channel
* 0 and the phonecall model to channel 1).
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/multichannel
*/
multichannel?: boolean;
/**
* Maximum number of transcript alternatives to return. Just like a human listener,
* Deepgram can provide multiple possible interpretations of what it hears.
* @default 1
*/
alternatives?: number;
/**
* Indicates whether to convert numbers from written format (e.g., one) to
* numerical format (e.g., 1). Deepgram can format numbers up to 999,999.
* @remarks Converted numbers do not include punctuation. For example,
* 999,999 would be transcribed as 999999.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/numerals
*/
numerals?: boolean;
/**
* Terms or phrases to search for in the submitted audio. Deepgram searches
* for acoustic patterns in audio rather than text patterns in transcripts
* because we have noticed that acoustic pattern matching is more performant.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/search
*/
search?: Array<string>;
/**
* Callback URL to provide if you would like your submitted audio to be
* processed asynchronously. When passed, Deepgram will immediately respond
* with a request_id. When it has finished analyzing the audio, it will send
* a POST request to the provided URL with an appropriate HTTP status code.
* @remarks You may embed basic authentication credentials in the callback URL.
* Only ports 80, 443, 8080, and 8443 can be used for callbacks.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/callback
*/
callback?: string;
/**
* Keywords to which the model should pay particular attention to boosting
* or suppressing to help it understand context. Just like a human listener,
* Deepgram can better understand mumbled, distorted, or otherwise
* hard-to-decipher speech when it knows the context of the conversation.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeAudio/properties/keywords
*/
keywords?: Array<string>;
/**
* Indicates whether the streaming endpoint should send you updates to its
* transcription as more audio becomes available. By default, the streaming
* endpoint returns regular updates, which means transcription results will
* likely change for a period of time. You can avoid receiving these updates
* by setting this flag to false.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeStreamingAudio/properties/interim_results
*/
interim_results?: boolean;
/**
* Indicates whether Deepgram will detect whether a speaker has finished
* speaking (or paused for a significant period of time, indicating the
* completion of an idea). When Deepgram detects an endpoint, it assumes
* that no additional data will improve its prediction, so it immediately
* finalizes the result for the processed time range and returns the
* transcript with a speech_final parameter set to true.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeStreamingAudio/properties/endpointing
*/
endpointing?: boolean;
/**
* Length of time in milliseconds of silence that voice activation detection
* (VAD) will use to detect that a speaker has finished speaking. Used when
* endpointing is enabled. Defaults to 10 ms. Deepgram customers may configure
* a value between 10 ms and 500 ms; on-premise customers may remove this
* restriction.
* @default 10
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeStreamingAudio/properties/vad_turnoff
*/
vad_turnoff?: number;
/**
* Expected encoding of the submitted streaming audio.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeStreamingAudio/properties/encoding
*/
encoding?: string;
/**
* Number of independent audio channels contained in submitted streaming
* audio. Only read when a value is provided for encoding.
* @default 1
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeStreamingAudio/properties/channels
*/
channels?: number;
/**
* Sample rate of submitted streaming audio. Required (and only read)
* when a value is provided for encoding.
* @see https://developers.deepgram.com/api-reference/speech-recognition-api#operation/transcribeStreamingAudio/properties/sample_rate
*/
sample_rate?: number;
}
{
channel_index: Array<number>;
duration: number;
start: number;
is_final: boolean;
speech_final: boolean;
channel: {
search?: Array<{
query: string;
hits: Array<{
confidence: number;
start: number;
end: number;
snippet: string;
}>
}>,
alternatives: Array<{
transcript: string;
confidence: number;
words: Array<{
word: string;
start: number;
end: number;
confidence: number;
punctuated_word?: string;
}>
}>
}
};
Retrieve all projects
const projects = await deepgram.projects.list();
{
projects: [
{
id: string,
name: string,
},
],
}
Retrieves all project based on the provided project id.
const project = await deepgram.projects.get(PROJECT_ID);
{
id: string,
name: string,
}
Updates a project based on a provided project object. This object must contain
project_id and name properties.
const updateResponse = await deepgram.projects.update(project);
{
message: string;
}
Retrieves all keys for a given project.
const response = await deepgram.keys.list(PROJECT_ID);
{
api_keys: [
{
api_key_id: string,
comment: string,
created: string,
scopes: Array<string>
},
];
}
Create a new API key for a project using the keys.create method
with a name for the key.
const response = await deepgram.keys.create(PROJECT_ID, COMMENT_FOR_KEY);
{
api_key_id: string,
key: string,
comment: string,
created: string,
scopes: Array<string>
}
Delete an existing API key using the keys.delete method with the key to
delete.
await deepgram.keys.delete(PROJECT_ID, KEY_ID);
Retrieves transcription requests for a project based on the provided options.
const response = await deepgram.usage.listRequests(PROJECT_ID, {
limit: 10,
// other options are available
});
{
// The time to retrieve requests made since
// Example: "2020-01-01T00:00:00+00:00"
start?: string,
// The time to retrieve requests made until
// Example: "2021-01-01T00:00:00+00:00"
end?: string,
// Page of requests to return
// Defaults to 0
page?: number,
// Number of requests to return per page
// Defaults to 10. Maximum of 100
limit?: number,
// Filter by succeeded or failed requests
// By default, all requests are returned
status?: 'succeeded' | 'failed'
}
{
page: number,
limit: number,
requests?: [
{
request_id: string;
created: string;
path: string;
accessor: string;
response?: {
details: {
usd: number;
duration: number;
total_audio: number;
channels: number;
streams: number;
model: string;
method: string;
tags: Array<string>;
features: Array<string>;
config: {
multichannel?: boolean;
interim_results?: boolean;
punctuate?: boolean;
ner?: boolean;
utterances?: boolean;
replace?: boolean;
profanity_filter?: boolean;
keywords?: boolean;
sentiment?: boolean;
diarize?: boolean;
detect_language?: boolean;
search?: boolean;
redact?: boolean;
alternatives?: boolean;
numerals?: boolean;
};
}
}, ||
{
message?: string;
},
callback?: {
code: number;
completed: string;
},
},
];
}
Retrieves a specific transcription request for a project based on the provided
projectId and requestId.
const response = await deepgram.usage.getRequest(PROJECT_ID, REQUEST_ID);
{
request_id: string;
created: string;
path: string;
accessor: string;
response?: {
details: {
usd: number;
duration: number;
total_audio: number;
channels: number;
streams: number;
model: string;
method: string;
tags: Array<string>;
features: Array<string>;
config: {
multichannel?: boolean;
interim_results?: boolean;
punctuate?: boolean;
ner?: boolean;
utterances?: boolean;
replace?: boolean;
profanity_filter?: boolean;
keywords?: boolean;
sentiment?: boolean;
diarize?: boolean;
detect_language?: boolean;
search?: boolean;
redact?: boolean;
alternatives?: boolean;
numerals?: boolean;
};
}
}, ||
{
message?: string;
},
callback?: {
code: number;
completed: string;
}
}
Retrieves aggregated usage data for a project based on the provided options.
const response = await deepgram.usage.getUsage(PROJECT_ID, {
start: "2020-01-01T00:00:00+00:00",
// other options are available
});
{
// The time to retrieve requests made since
// Example: "2020-01-01T00:00:00+00:00"
start?: string,
// The time to retrieve requests made until
// Example: "2021-01-01T00:00:00+00:00"
end?: string,
// Specific identifer for a request
accessor?: string,
// Array of tags used in requests
tag?: Array<string>,
// Filter requests by method
method?: "sync" | "async" | "streaming",
// Filter requests by model used
model?: string,
// Filter only requests using multichannel feature
multichannel?: boolean,
// Filter only requests using interim results feature
interim_results?: boolean,
// Filter only requests using the punctuation feature
punctuate?: boolean,
// Filter only requests using ner feature
ner?: boolean,
// Filter only requests using utterances feature
utterances?: boolean,
// Filter only requests using replace feature
replace?: boolean,
// Filter only requests using profanity_filter feature
profanity_filter?: boolean,
// Filter only requests using keywords feature
keywords?: boolean,
// Filter only requests using sentiment feature
sentiment?: boolean,
// Filter only requests using diarization feature
diarize?: boolean,
// Filter only requests using detect_language feature
detect_language?: boolean,
// Filter only requests using search feature
search?: boolean,
// Filter only requests using redact feature
redact?: boolean,
// Filter only requests using alternatives feature
alternatives?: boolean,
// Filter only requests using numerals feature
numerals?: boolean
}
{
start: string,
end: string,
resolution: {
units: string,
amount: number
};
results: [
{
start: string,
end: string,
hours: number,
requests: number
}
];
}
Retrieves features used by the provided projectId based on the provided options.
const response = await deepgram.usage.getUsage(PROJECT_ID, {
start: "2020-01-01T00:00:00+00:00",
// other options are available
});
{
// The time to retrieve requests made since
// Example: "2020-01-01T00:00:00+00:00"
start?: string,
// The time to retrieve requests made until
// Example: "2021-01-01T00:00:00+00:00"
end?: string
}
{
tags: Array<string>,
models: Array<string>,
processing_methods: Array<string>,
languages: Array<string>,
features: Array<string>
}
To run the sample code, first run the following in your terminal:
npm install
npm build
Then update the config object located at the top of the index.js
file in the sample folder.
const config = {
deepgramApiKey: "YOUR_DEEPGRAM_API_KEY",
urlToFile:
"https://static.deepgram.com/examples/Bueller-Life-moves-pretty-fast.wav",
};
Finally, run the sample code using the following command in your terminal:
node sample/index.js
The sample demonstrates the following uses:
Interested in contributing? We ❤️ pull requests!
To make sure our community is safe for all, be sure to review and agree to our Code of Conduct. Then see the Contribution guidelines for more information.
We love to hear from you so if you have questions, comments or find a bug in the project, let us know! You can either:
Check out the Developer Documentation at https://developers.deepgram.com/
FAQs
Isomorphic Javascript client for Deepgram
The npm package @deepgram/sdk receives a total of 45,390 weekly downloads. As such, @deepgram/sdk popularity was classified as popular.
We found that @deepgram/sdk demonstrated a healthy version release cadence and project activity because the last version was released less than a year ago. It has 1 open source maintainer collaborating on the project.
Did you know?
Socket for GitHub automatically highlights issues in each pull request and monitors the health of all your open source dependencies. Discover the contents of your packages and block harmful activity before you install or update your dependencies.
Security News
Company News
Socket is joining TC54 to help develop standards for software supply chain security, contributing to the evolution of SBOMs, CycloneDX, and Package URL specifications.
Security News
PyPI now allows maintainers to archive projects, improving security and helping users make informed decisions about their dependencies.
Research
Security News
Malicious npm package postcss-optimizer delivers BeaverTail malware, targeting developer systems; similarities to past campaigns suggest a North Korean connection.