Voice Assistants

Hume AI

APIs, datasets, and models to add emotional intelligence to voice AI applications

hume.ai

Verified API available ~4.2k monthly visits

Quick facts
What is it	APIs, datasets, and models to add emotional intelligence to voice AI applications
Pricing	Paid — from $3/mo
Free tier	Yes
Platform	Web Application
API	Yes
Best for	audiobook narration, podcast production
Domain registered	2020

Data updated Aug. 1, 2026

What does Hume AI do?

Hume AI is a research lab that sells tools to make voice AI more emotionally aware. Instead of just processing words, its models can detect and express nuances like frustration, excitement, or hesitation. The company offers APIs for running human feedback studies, curated speech datasets, and both open-source and proprietary voice models.

The Human Feedback API gives you pre-built survey templates designed by scientists to capture preference data on audio quality, listenability, and emotional alignment. You can set up studies and get results in hours, not weeks. For training, Hume's datasets cover over 48 emotions across 50+ languages, with annotations for conversational dynamics, emotional reproduction, and voice realism. On the model side, they have TADA (an open-source text-to-speech system that streams text and audio together to reduce latency and hallucinations), Octave (a closed-source TTS with voice cloning and modulation), and EVI (a speech-to-speech model that handles interruptions and backchanneling).

Developers building voice assistants, conversational AI for customer service, or interactive characters for games and entertainment will find these tools useful. Researchers in affective computing can use the datasets and open-source models to advance their own work. Hume positions itself as a research-first company, which means the APIs and models are grounded in academic studies, not just hype.

Text-to-Speech Voice Assistants

#ai agents#ai chatbots#art generators#audio generators#best ai 3d generators#best ai art generators#best ai chatbots#best ai image generators#best ai text generators#customer support#image generators#productivity tools#research#text generators#video tools

Key features

What makes it stand out

Human Feedback API with science-backed survey templates for fast, reliable preference data

Curated speech datasets annotated across 48 emotions and 50+ languages

Open-source TTS model (TADA) that reduces latency and hallucinations by streaming text and audio together

Closed-source TTS (Octave) with voice cloning, modulation, and conversion

Speech-to-speech model (EVI) that supports interruptions, backchanneling, and expressive instruction following

Who is Hume AI for?

Who benefits most from this tool

audiobook narration

podcast production

video voiceover creation

Pricing

Free tier available — start without a credit card

Free

15 rpm
10,000 monthly characters
5 monthly evi minutes
1 concurrent connections
Text-to-speech (Octave 1, Octave 2)
Speech-to-speech (EVI 3, EVI 4 mini)
Voice cloning (create only)
Discord support

Starter

$3.0/month

15 rpm
20 projects
30,000 monthly characters
40 monthly evi minutes
5 concurrent connections
Text-to-speech (Octave 1, Octave 2)
Speech-to-speech (EVI 3, EVI 4 mini)
Voice cloning (create only)
20 projects
Discord support

Creator

$7.0/month

75 rpm
1,000 projects
140,000 monthly characters
200 monthly evi minutes
0.07 evi 3 cost per minute
5 concurrent connections
0.15 additional characters per 1k
Text-to-speech (Octave 1, Octave 2)
Speech-to-speech (EVI 3, EVI 4 mini)
Unlimited voice cloning
Commercial license
Discord support

Pro

$70.0/month

75 rpm
3,000 projects
1,000,000 monthly characters
1,200 monthly evi minutes
0.06 evi 3 cost per minute
10 concurrent connections
0.12 additional characters per 1k
Text-to-speech (Octave 1, Octave 2)
Speech-to-speech (EVI 3, EVI 4 mini)
Unlimited voice cloning
Commercial license
Discord support

Scale

$200.0/month

150 rpm
10,000 projects
3 team seats
3,300,000 monthly characters
5,000 monthly evi minutes
0.05 evi 3 cost per minute
20 concurrent connections
0.1 additional characters per 1k
Text-to-speech (Octave 1, Octave 2)
Speech-to-speech (EVI 3, EVI 4 mini)
Unlimited voice cloning
Commercial license
Organization features
3 team seats
Discord support

Business

$500.0/month

225 rpm
20,000 projects
5 team seats
10,000,000 monthly characters
12,500 monthly evi minutes
0.04 evi 3 cost per minute
30 concurrent connections
0.05 additional characters per 1k
Text-to-speech (Octave 1, Octave 2)
Speech-to-speech (EVI 3, EVI 4 mini)
Unlimited voice cloning
Commercial license
Organization features
5 team seats
Discord support

Enterprise

Custom

custom rpm
unlimited projects
unlimited team seats
unlimited monthly characters
unlimited monthly evi minutes
custom evi 3 cost per minute
unlimited concurrent connections
custom additional characters per 1k
Text-to-speech (Octave 1, Octave 2)
Speech-to-speech (EVI 3, EVI 4 mini)
Unlimited voice cloning (create, use and access via API)
Commercial license
Organization features
Unlimited team seats
Custom support

Trust & presence

Domain Domain registered 2020

Gallery

Click any image to enlarge

Alternatives in Voice Assistants

Sesame Voice Assistants

AI voice companion that understands conversational context, tone, and emotion for natural dialogue

ifttt

Toum AI Voice Assistants

Voice AI platform for multilingual markets — build, deploy, and manage voice agents that speak local languages and dialects

Deepgram Voice AI Voice Assistants

APIs for speech-to-text, text-to-speech, and voice agents — add voice AI to your applications.

Sun Voice Assistants

Realtime voice AI model that handles multi-speaker conversations, interruptions, and follow-ups without wake words

audeering.com Voice Assistants

AI voice analysis platform that detects emotion, health biomarkers, and speaker attributes from vocal expressions.

Tunk Voice Assistants

Voice AI platform for creating human-like AI agents that handle inbound/outbound calls and speech-to-text in 50+ languages.

Deepgram Voice Assistants

Enterprise Voice AI platform designed for developers building voice-first products using speech-to-text, text-to-speech, or speech-to-speech APIs. Over 200,000 developers build with Deepgram's voice-native foundational models, accessed via APIs or self-managed software. Start building with $200 in free credits!

zapier · make+2 Top 100k site

Thinnest AI Voice Assistants

Build and deploy voice AI agents that make human-like calls in 100+ languages, starting at ₹1.5/min.

Similar tools

Humii Life Assistant

AI bedtime companion — pick a creator's voice, share your day, and get a personalized audio story to help you fall asleep

Hedy AI Meeting Assistant

AI conversation coach — get real-time feedback on your speaking in meetings, interviews, and classes.

zapier · integrately

Theysaid Ux Research

AI user testing and feedback platform — run interactive surveys, interviews, and tests with real-time AI analysis.

Humy.ai education

AI-powered historical figures for education — chat with 1,200+ personas to make history lessons interactive and engaging.