Loading models...
Audio Intelligence
Audio intelligence features are powered by task-specific language models. When an audio intelligence feature is enabled, Punctuation will also be enabled by default.summarize=falseIndicates whether Deepgram will provide summaries for sections of content.
topics=falseIndicates whether Deepgram will identify and extract key topics for sections of content.
intents=falseIndicates whether Deepgram will recognize intents.
detect_entities=falseIndicates whether Deepgram will identify and extract key entities for sections of content.
sentiment=falseIdentifies sentiment of positive, neutral, or negative, in addition to a sentiment score. Provides sentiment on the word level, as well as for sentences, paragraphs and segments.
Transcription
Transcription features work with audio input. Selection varies between pre-recorded and live audio requests.smart_format=trueSmart Format improves readability by applying additional formatting. When enabled, punctuation and paragraph breaks will be applied as well as formatting of other entities, such as dates, times, and numbers.
keyterm=TERM_OR_PHRASEKeyterms that the model should pay particular attention to (e.g. proper nouns or product names). Supports up to 100 keyterms in a single request.
punctuate=falseIndicates whether to add punctuation and capitalization to the transcript.
paragraphs=falseIndicates whether Deepgram will split audio into paragraphs to improve transcript readability. When paragraphs is set to true, punctuate will also be set to true.
utterances=falseSegments speech into meaningful semantic units. By default, when utterances is enabled, it starts a new utterance after 0.8s of silence. You can customize the length of time used to determine where to split utterances with the utt_split parameter.
profanity_filter=falseIndicates whether to remove profanity from the transcript.
redact=OPTIONIndicates whether to redact sensitive information, replacing redacted content with asterisks (*).
replace=FIND%3AREPLACETerms or phrases to search for in the submitted audio and replace. The Find value must be lowercase and will match words from the word field. The Replace value can be any case.
search=TERM_OR_PHRASETerms 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.
keywords=KEYWORD%3AINTENSIFIERKeywords to which the model should pay particular attention to boosting or suppressing to help it understand context. Intensifier indicates how much you want to boost it. The default Intensifier is one (1). An Intensifier of two (2) equates to two boosts multiplied in a row, whereas zero (0) is equivalent to not specifying a keywords parameter at all.
diarize=falseIndicates whether to recognize speaker changes.
filler_words=falseIndicates whether Deepgram will transcribe disfluencies in your audio, like "uh" and "um".
If you want to test out features with text input, switch to “Text Intelligence” in the top bar.
We are improving the Playground and appreciate your opinion. Please let us know how we can improve.
Playground
Deepgram's Playground is best experienced on large screens. Try widening your browser window.
Learn more about us at deepgram.com, our documentation, by signing up for free, or by meeting our Voice Agent API.