Skip to content
AI.info

tools

Empathic Voice Interface (EVI)

EVI is Hume's real-time voice AI for building conversations that respond to vocal expression and emotional context.

Empathic Voice Interface (EVI)

Empathic Voice Interface measures vocal expression, processes speech and text, and produces streamed speech responses. Developers can connect it to their own language model, configure voices, add tools, inject context, and access transcripts with emotion data.

EVI is designed for embedded voice applications such as assistants, customer service, accessibility tools, games, and robotics. It is a developer API and SDK rather than a standalone consumer chat app. Plans include monthly EVI minutes; additional usage and enterprise limits cost extra.

Features

  • Measures vocal expression and responds with emotionally attuned speech
  • Supports EVI 3 and EVI 4-mini configurations
  • Connects to external language models or Hume's speech-language models
  • Streams audio responses in real time with about 300ms time to first byte
  • Supports expressive voice selection, voice design, and voice cloning
  • Provides interruption, pause, resume, and chat-history controls
  • Supports dynamic variables, context injection, and tool use
  • Reconstructs complete audio from conversations on demand

Use cases

  • Build voice-first AI assistants with natural turn-taking
  • Add emotionally responsive voice support to customer-service systems
  • Create accessible voice interfaces for applications and services
  • Develop conversational characters for games, VR, and immersive experiences
  • Connect voice agents to appointments, orders, and other APIs

Pros

    Cons

      Latest updates

      • TTS and EVI API sunset notice

        Hume is sunsetting the TTS and EVI APIs. Access ends November 13, 2026 at 12:01 a.m. EST.

      • TTS API additions

        Added an experimental temperature parameter to TTS endpoints to control sampling temperature for speech generation.

      • EVI API additions

        Added configurable turn detection and interruption settings to EVI configs.

      • EVI API additions

        Added support for new supplemental LLM models and zero prompt expansion.

      • TTS API bug fixes

        Fixed a bug where duplicate interleaved audio was included in TTS audio output.

      Capabilities

      • Text to speech — “Streamed speech generation via our speech-language model.” source
      • Voice cloning — “Design a voice, clone an existing one, or select one from Hume’s extensive Voice Library .” source
      • Speech to text — “Fast and accurate ASR returns a full transcript of the conversation, with Hume’s expression measures tied to each sentence.” source
      • Many languages — “EVI 4-mini supports English, Japanese, Korean, Spanish, French, Portuguese, Italian, German, Russian, Hindi, Arabic.” source
      • API — “The EVI API supports thousands of concurrent sessions.” source

      Pricing

      Starting price
      $3/mo
      Prices checked
      2026-10-24

      Free

      Free

      • 10,000 monthly included characters
      • 5 minutes monthly EVI usage included
      • 15 requests per minute
      • 1 concurrent connection
      • Unlimited voice cloning (create and use)

      Starter

      • $3 / month
      • 30,000 monthly included characters
      • 40 minutes monthly EVI usage included
      • 15 requests per minute
      • 5 concurrent connections
      • Unlimited voice cloning (create and use)

      Creator

      • $14 / month
      • 140,000 monthly included characters
      • 200 minutes monthly EVI usage included
      • 75 requests per minute
      • 5 concurrent connections
      • Unlimited voice cloning (create and use)

      Pro

      • $70 / month
      • 1,000,000 monthly included characters
      • 1,200 minutes monthly EVI usage included
      • 75 requests per minute
      • 10 concurrent connections
      • Unlimited voice cloning (create and use)

      Scale

      • $200 / month
      • 3,300,000 monthly included characters
      • 5,000 minutes monthly EVI usage included
      • 150 requests per minute
      • 20 concurrent connections
      • 3 team seats
      • Unlimited voice cloning (create and use)

      Business

      • $500 / month
      • 10,000,000 monthly included characters
      • 12,500 minutes monthly EVI usage included
      • 225 requests per minute
      • 30 concurrent connections
      • 5 team seats
      • Unlimited voice cloning (create and use)

      Enterprise

      Price on request

      • As much monthly included characters as you need
      • As much monthly EVI usage as you need
      • As many concurrent connections as you need
      • Unlimited voice cloning (create, use and access via API)
      • Unlimited team seats
      • Slack support
      Official website