Hume AI EVI

Hume AI's Empathic Voice Interface provides realtime speech-to-speech conversations with expressive voices, interruption handling, tools, custom language models, and web or app SDKs. It is best treated as a developer platform whose expression and prosody signals require careful validation, not as a system that can reliably know a caller's emotions.

Editorial research · Reviewed Jul 10, 2026

Test the agent live

Official Hume conversational AI demo; browser microphone access is required.

Open the live demo

Quick facts

Platform type

Voice Agent Platform

Inbound calls

Outbound calls

Human handoff

Yes

Setup difficulty

Technical

Pricing model

$0.04/min for listed EVI usage; external services may add cost

Developer friendly

Yes

Pricing details

Hume's pricing page lists EVI usage at $0.04/min on the shown plan, with included minutes by subscription tier and custom enterprise pricing. External LLMs, telephony, and other services may be additional.

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

$160

4,000 min / month

$1,920 / year

Integrations

Web SDKReactAPILiveKitExternal LLMsCustom tools

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

Avoid treating inferred expression as a diagnosis or ground truth. Obtain microphone and recording consent, disclose AI interaction, scope API keys and tools, and review voice-cloning rights and retention.