Deepgram

Deepgram offers speech-to-text, text-to-speech, audio intelligence, and a Voice Agent API that combines STT, LLM orchestration, TTS, turn-taking, interruption handling, and function calling in a single real-time voice stack. It is a strong fit for developers and enterprises that want voice-agent infrastructure rather than a packaged receptionist.

Test the agent live

Talk to the Voice Agent API in Deepgram's browser playground.

Open the live demo

Quick facts

Platform type

Developer API

Inbound calls

Yes

Outbound calls

Yes

Human handoff

Yes

Setup difficulty

Technical

Pricing model

$4.50/hour ($0.075/min) for Voice Agent API

Developer friendly

Yes

Pricing details

Deepgram lists Voice Agent API pricing at $4.50/hour ($0.075/min) when using Deepgram's full stack, with built-in rate reductions when bringing your own models. Deepgram also offers managed, dedicated single-tenant, VPC, and self-hosted deployment options. Verify with vendor.

Cost estimator

A rough monthly usage cost from your call volume. Always confirm rates with the vendor.

Calls / month

Avg length

min

Rate / min

Estimated monthly

$300

4,000 min / month

$3,600 / year

Same volume across platforms

4,000 min/mo · usage only

Deepgram (this)$300
AssemblyAI$300
Vapi$920

Comparable platforms priced at your volume using vendor list rates. Verify rates and add any base or provider fees. Compare all platforms at your volume →

Integrations

WebSocket APICustom APIFunction callingTwilioBYO LLMBYO TTS

Data and permissions

Voice agents may talk to customers, access records, record calls, and trigger external actions.

Can read customer data

Can write or update external systems

Can send messages

Can trigger workflows

Requires API keys

Records calls

Human approval available

Review audio and transcript retention, API key handling, tool-call permissions, deployment mode, HIPAA/GDPR needs, regional data residency, and any self-hosted or VPC requirements.