API Rate Limits

Understand the different service limits of Deepgram’s APIs.

Pay as You Go / Growth

Limits to consider if you use the Pay as You Go or Growth plans with Deepgram.

Voice Agent

APIConnection Limits
Voice Agent APIUp to 5 concurrent connections

Speech to Text

If multiple services are used in one API call (e.g Speech to Text + Sentiment Analysis), the lower of the rate limits is applied.

ModelService Limit
Nova-2Pre-Recorded Up to 100 concurrent requests
Streaming Up to 100 concurrent requests
NovaPre-Recorded Up to 100 concurrent requests
Streaming Up to 100 concurrent requests
EnhancedPre-Recorded Up to 100 concurrent requests
Streaming Up to 100 concurrent requests
BasePre-Recorded Up to 100 concurrent requests
Streaming Up to 100 concurrent requests
Whisper CloudPre-Recorded Up to 5 concurrent requests

Text to Speech REST

ModelService Limit
AuraPay as You Go: Up to 480 requests / min
AuraGrowth: Up to 720 requests / min

Text to Speech Streaming

ModelService Limit
AuraPay as You Go: Up to 40 concurrent requests
AuraGrowth: Up to 80 concurrent requests

Audio Intelligence

If you include Audio Intelligence features in requests to /listen, you will be subject to the service limits noted in the table below.

ModelService Limit
Intent RecognitionUp to 10 concurrent requests
Entity DetectionUp to 5 concurrent requests
Sentiment AnalysisUp to 10 concurrent requests
SummarizationUp to 10 concurrent requests
Topic DetectionUp to 10 concurrent requests

Text Intelligence

ModelService Limit
Intent RecognitionUp to 10 concurrent requests
Sentiment AnalysisUp to 10 concurrent requests
SummarizationUp to 10 concurrent requests
Topic DetectionUp to 10 concurrent requests

Enterprise

Starting limits to consider if you have an Enterprise Contract with Deepgram.

New and existing Enterprise customers can request a Service Limit increase by discussing your needs with the Deepgram Sales Team.

Voice Agent

APIConnection Limits
Voice Agent APIUp to 5 concurrent connections

Speech to Text

If multiple services are used in one API call (e.g Speech to Text + Sentiment Analysis), the lower of the rate limits is applied.

ModelService Limit
Nova-2Pre-Recorded Starting at 100 concurrent requests
Streaming Starting at 100 concurrent requests
NovaPre-Recorded Starting at 100 concurrent requests
Streaming Starting at 100 concurrent requests
EnhancedPre-Recorded Starting at 100 concurrent requests
Streaming Starting at 100 concurrent requests
BasePre-Recorded Starting at 100 concurrent requests
Streaming Starting at 100 concurrent requests
Whisper CloudPre-Recorded Starting at 15 concurrent requests

Text to Speech REST

ModelService LimitFeedback
AuraStarting at 2400 requests / minShare your feedback on our TTS rate limits

Text to Speech Streaming

ModelService LimitFeedback
AuraStarting at 150 concurrent requestsShare your feedback on our TTS rate limits

Audio Intelligence

If you include Audio Intelligence features in requests to /listen, you will be subject to the service limits noted in the table below.

ModelService Limit
Intent RecognitionStarting at 10 concurrent requests
Entity DetectionStarting at 10 concurrent requests
Sentiment AnalysisStarting at 10 concurrent requests
SummarizationStarting at 10 concurrent requests
Topic DetectionStarting at 10 concurrent requests

Text Intelligence

ModelService Limit
Intent RecognitionStarting at 10 concurrent requests
Sentiment AnalysisStarting at 10 concurrent requests
SummarizationStarting at 10 concurrent requests
Topic DetectionStarting at 10 concurrent requests

The error 429: Too Many Requests is returned when your project has more concurrent requests than the rate limits allow. To learn more about this error please see our Documentation on Errors.

Built with