Deepgram | Nova-3 | Speech to Text Pro
Advanced speech-to-text transcription powered by Deepgram Nova-3 with speaker diarization, named entity detection, intent classification, sentiment analysis, topic detection, and automatic summarization.
Quick Start
Requires an each::labs API key. Get one at eachlabs.ai.
curl -X POST https://api.eachlabs.ai/v1/prediction \
-H "Content-Type: application/json" \
-H "X-API-Key: $EACHLABS_API_KEY" \
-d '{
"model": "deepgram-nova-3-speech-to-text-pro",
"version": "0.0.1",
"input": {
"media_url": "https://example.com/meeting-recording.mp3",
"model": "nova-3",
"language_code": "en",
"diarize": true,
"summarize": true,
"sentiment": true
}
}'
Parameters
| Parameter |
Type |
Default |
Description |
| detect_entities |
boolean |
true |
Extract named entities: persons, organizations, locations, dates, ordinals. |
| diarize |
boolean |
true |
Identify different speakers. Each word includes speaker ID and confidence score. |
| filler_words |
boolean |
true |
Include filler words (uh, um) in transcript. |
| intents |
boolean |
true |
Detect speaker intents. Returns intent labels with confidence scores per segment. |
| language_code |
string |
auto |
Language of the audio (BCP-47). auto: automatic detection. multi: multilingual. enum: auto, multi, tr, ur, en, nl, uk, es, ar, de, fr, it, ja, ko, pt, ru, zh, hi, and more |
| media_url |
string |
|
Audio file URL to transcribe. Supports mp3, wav, m4a, flac, ogg, webm, mp4. |
| model |
string |
nova-3 |
Deepgram model. nova-3: latest, best accuracy. nova-2: previous generation. enum: nova-3, nova-2 |
| multichannel |
boolean |
false |
Transcribe each audio channel independently. Max 5 channels. |
| numerals |
boolean |
false |
Convert spoken numbers to digits (e.g., three hundred to 300). |
| paragraphs |
boolean |
true |
Split transcript into paragraphs with sentences. Requires punctuate=true. |
| profanity_filter |
boolean |
false |
Replace profanity with asterisks in transcript. |
| punctuate |
boolean |
true |
Add punctuation and capitalization to transcript. |
| redact |
string |
false |
Redact sensitive info. false: none. pci: credit cards. pii: SSN/phone/email. enum: false, pci, pii, numbers |
| sentiment |
boolean |
true |
Analyze sentiment per segment and overall. |
| smart_format |
boolean |
true |
Auto-format currency, phone numbers, emails, dates for readability. |
| summarize |
boolean |
true |
Generate a short text summary of the transcript content. |
| topics |
boolean |
true |
Detect topics discussed. Returns topic labels with confidence scores. |
| utt_split |
number |
0.8 |
Seconds of silence to split utterances. Lower = more splits. |
| utterances |
boolean |
true |
Segment speech into semantic units with start/end times and speaker info. |
Examples
Full meeting transcription with all features:
curl -X POST https://api.eachlabs.ai/v1/prediction \
-H "Content-Type: application/json" \
-H "X-API-Key: $EACHLABS_API_KEY" \
-d '{
"model": "deepgram-nova-3-speech-to-text-pro",
"version": "0.0.1",
"input": {
"media_url": "https://example.com/team-meeting.mp3",
"model": "nova-3",
"language_code": "en",
"diarize": true,
"summarize": true,
"sentiment": true,
"detect_entities": true,
"intents": true,
"topics": true,
"paragraphs": true
}
}'
Simple transcription with PII redaction:
curl -X POST https://api.eachlabs.ai/v1/prediction \
-H "Content-Type: application/json" \
-H "X-API-Key: $EACHLABS_API_KEY" \
-d '{
"model": "deepgram-nova-3-speech-to-text-pro",
"version": "0.0.1",
"input": {
"media_url": "https://example.com/customer-call.wav",
"model": "nova-3",
"redact": "pii",
"profanity_filter": true,
"numerals": true
}
}'
Multilingual transcription:
curl -X POST https://api.eachlabs.ai/v1/prediction \
-H "Content-Type: application/json" \
-H "X-API-Key: $EACHLABS_API_KEY" \
-d '{
"model": "deepgram-nova-3-speech-to-text-pro",
"version": "0.0.1",
"input": {
"media_url": "https://example.com/multilingual-podcast.mp3",
"language_code": "multi",
"diarize": true,
"filler_words": false,
"smart_format": true
}
}'
Related Models
Documentation
1---2name: deepgram-nova-3-speech-to-text-pro3description: Deepgram | Nova-3 | Speech to Text Pro. Advanced speech transcription with diarization, sentiment, entities, intents, and summarization. Triggers: transcribe, speech to text, stt, deepgram, nova, diarize, sentiment4---56# Deepgram | Nova-3 | Speech to Text Pro78Advanced speech-to-text transcription powered by Deepgram Nova-3 with speaker diarization, named entity detection, intent classification, sentiment analysis, topic detection, and automatic summarization.910## Quick Start1112> Requires an each::labs API key. Get one at [eachlabs.ai](https://eachlabs.ai).1314```bash15curl -X POST https://api.eachlabs.ai/v1/prediction \16 -H "Content-Type: application/json" \17 -H "X-API-Key: $EACHLABS_API_KEY" \18 -d '{19 "model": "deepgram-nova-3-speech-to-text-pro",20 "version": "0.0.1",21 "input": {22 "media_url": "https://example.com/meeting-recording.mp3",23 "model": "nova-3",24 "language_code": "en",25 "diarize": true,26 "summarize": true,27 "sentiment": true28 }29 }'30```3132## Parameters3334| Parameter | Type | Default | Description |35|-----------|------|---------|-------------|36| detect_entities | boolean | true | Extract named entities: persons, organizations, locations, dates, ordinals. |37| diarize | boolean | true | Identify different speakers. Each word includes speaker ID and confidence score. |38| filler_words | boolean | true | Include filler words (uh, um) in transcript. |39| intents | boolean | true | Detect speaker intents. Returns intent labels with confidence scores per segment. |40| language_code | string | auto | Language of the audio (BCP-47). auto: automatic detection. multi: multilingual. enum: auto, multi, tr, ur, en, nl, uk, es, ar, de, fr, it, ja, ko, pt, ru, zh, hi, and more |41| media_url | string | | Audio file URL to transcribe. Supports mp3, wav, m4a, flac, ogg, webm, mp4. |42| model | string | nova-3 | Deepgram model. nova-3: latest, best accuracy. nova-2: previous generation. enum: nova-3, nova-2 |43| multichannel | boolean | false | Transcribe each audio channel independently. Max 5 channels. |44| numerals | boolean | false | Convert spoken numbers to digits (e.g., three hundred to 300). |45| paragraphs | boolean | true | Split transcript into paragraphs with sentences. Requires punctuate=true. |46| profanity_filter | boolean | false | Replace profanity with asterisks in transcript. |47| punctuate | boolean | true | Add punctuation and capitalization to transcript. |48| redact | string | false | Redact sensitive info. false: none. pci: credit cards. pii: SSN/phone/email. enum: false, pci, pii, numbers |49| sentiment | boolean | true | Analyze sentiment per segment and overall. |50| smart_format | boolean | true | Auto-format currency, phone numbers, emails, dates for readability. |51| summarize | boolean | true | Generate a short text summary of the transcript content. |52| topics | boolean | true | Detect topics discussed. Returns topic labels with confidence scores. |53| utt_split | number | 0.8 | Seconds of silence to split utterances. Lower = more splits. |54| utterances | boolean | true | Segment speech into semantic units with start/end times and speaker info. |5556## Examples5758**Full meeting transcription with all features:**59```bash60curl -X POST https://api.eachlabs.ai/v1/prediction \61 -H "Content-Type: application/json" \62 -H "X-API-Key: $EACHLABS_API_KEY" \63 -d '{64 "model": "deepgram-nova-3-speech-to-text-pro",65 "version": "0.0.1",66 "input": {67 "media_url": "https://example.com/team-meeting.mp3",68 "model": "nova-3",69 "language_code": "en",70 "diarize": true,71 "summarize": true,72 "sentiment": true,73 "detect_entities": true,74 "intents": true,75 "topics": true,76 "paragraphs": true77 }78 }'79```8081**Simple transcription with PII redaction:**82```bash83curl -X POST https://api.eachlabs.ai/v1/prediction \84 -H "Content-Type: application/json" \85 -H "X-API-Key: $EACHLABS_API_KEY" \86 -d '{87 "model": "deepgram-nova-3-speech-to-text-pro",88 "version": "0.0.1",89 "input": {90 "media_url": "https://example.com/customer-call.wav",91 "model": "nova-3",92 "redact": "pii",93 "profanity_filter": true,94 "numerals": true95 }96 }'97```9899**Multilingual transcription:**100```bash101curl -X POST https://api.eachlabs.ai/v1/prediction \102 -H "Content-Type: application/json" \103 -H "X-API-Key: $EACHLABS_API_KEY" \104 -d '{105 "model": "deepgram-nova-3-speech-to-text-pro",106 "version": "0.0.1",107 "input": {108 "media_url": "https://example.com/multilingual-podcast.mp3",109 "language_code": "multi",110 "diarize": true,111 "filler_words": false,112 "smart_format": true113 }114 }'115```116117## Related Models118119- [deepgram-nova-3-speech-to-text](../deepgram-nova-3-speech-to-text/) - Standard Deepgram Nova-3 transcription120- [elevenlabs-speech-to-text-scribe-v2](../elevenlabs-speech-to-text-scribe-v2/) - ElevenLabs speech to text121122## Documentation123124- [each::labs Docs](https://docs.eachlabs.ai)125- [API Reference](https://docs.eachlabs.ai/api/overview)
awesome-genmedia