Deepgram
App · Document Extraction · 1 action
Read a document with Deepgram's text analysis: summarize it, score its sentiment, and pull out the topics and intents behind it, including topics and intents of your own.
Analyze text
Run text through Deepgram's text analysis and get back a summary, the overall sentiment, the topics it covers, and the intents behind it. Turn on the parts you need: a support email, a meeting transcript, or a survey response comes back as fields a later step can act on.
Inputs
- Text
texttext - The text to analyze. Paste it in, or chain it from an earlier step. Leave blank if you are pointing at a file address instead.
- Text file address
urltext - The address (URL) of a plain-text file for Deepgram to fetch and analyze, instead of pasting the text in. Use one or the other, not both.
- Write a summary
summarizetrue/false - Have Deepgram write a short summary of the text.Default:
false. - Detect sentiment
sentimenttrue/false - Rate the text as positive, neutral, or negative: overall and passage by passage.Default:
false. - Detect topics
topicstrue/false - Find what the text is about, passage by passage.Default:
false. - Detect intents
intentstrue/false - Find what the writer or speaker is trying to do, such as asking for a refund or booking a demo.Default:
false. - Custom topics
custom_topiclist - Topics of your own to look for, such as 'pricing' or 'cancellation'. Up to 100. Needs 'Detect topics' turned on.
- Custom topic mode
custom_topic_modetext - Whether Deepgram may also report topics of its own alongside your custom ones, or stick to your list. Only applies when you list custom topics above.One of:
extended(Mine plus Deepgram's (recommended)),strict(Only mine). Default:extended. - Custom intents
custom_intentlist - Intents of your own to look for, such as 'request a refund'. Up to 100. Needs 'Detect intents' turned on.
- Custom intent mode
custom_intent_modetext - Whether Deepgram may also report intents of its own alongside your custom ones, or stick to your list. Only applies when you list custom intents above.One of:
extended(Mine plus Deepgram's (recommended)),strict(Only mine). Default:extended. - Language
languagetext - The language of the text, as a BCP-47 tag such as 'en'. Text analysis supports fewer languages than the rest of Deepgram, so leave it on English unless their docs list yours.Default:
en. - Usage tags
taglist - Labels attached to the request so you can pick it out on Deepgram's usage reports. They don't affect the analysis.
Outputs
intent_segmentslist- Which intents were found in which passage, with confidence scores.
intentslist- The distinct intents found anywhere in the text.
languagetext- The language Deepgram analyzed the text as.
metadataobject- Deepgram's own record of the request: when it ran, which models answered, and the tokens each one used.
request_idtext- Deepgram's ID for this request, to quote to their support if something looks wrong.
sentimenttext- The overall sentiment: 'positive', 'neutral', or 'negative'. Only when sentiment was asked for.
sentiment_scorenumber- The overall sentiment as a number from -1 (negative) to 1 (positive).
sentiment_segmentslist- The sentiment of each passage, in order.
summarytext- The written summary. Only when a summary was asked for.
topic_segmentslist- Which topics were found in which passage, with confidence scores.
topicslist- The distinct topics found anywhere in the text.
Example output
{
"intent_segments": [
{
"end_word": 414,
"intents": [
{
"confidence_score": 0.0038975573,
"intent": "Encourage podcasting"
}
],
"start_word": 354,
"text": "If you found this valuable, you can subscribe to the show on Spotify or your favorite podcast app."
}
],
"intents": [
"Encourage podcasting"
],
"language": "en",
"metadata": {
"created": "2026-07-29T23:47:44.674Z",
"language": "en",
"request_id": "d04af392-db11-4c1d-83e1-20e34f0b8999",
"summary_info": {
"input_tokens": 350,
"model_uuid": "a1b2c3d4-e5f6-7890-abcd-ef1234567890",
"output_tokens": 75
}
},
"request_id": "d04af392-db11-4c1d-83e1-20e34f0b8999",
"sentiment": "positive",
"sentiment_score": 0.5810185185185185,
"sentiment_segments": [
{
"end_word": 69,
"sentiment": "positive",
"sentiment_score": 0.5810546875,
"start_word": 0,
"text": "As much as it's worth celebrating the first spacewalk with an all-female team, many of us are looking forward to it just being normal."
}
],
"summary": "This transcript highlights the significance of the first all-female spacewalk and honors the women pioneers who paved the way.",
"topic_segments": [
{
"end_word": 69,
"start_word": 32,
"text": "I think if it signifies anything, it is to honor the women who came before us who were skilled and qualified and didn't get the same opportunities that we have today.",
"topics": [
{
"confidence_score": 0.91581345,
"topic": "Spacewalk"
}
]
}
],
"topics": [
"Spacewalk"
]
}