Hume AI Inc. · Tools

Hume AI

Visit website

At a glance

Price
from 3 $/mo
Vendor
Hume AI Inc.

Specifications & properties

Key decision factors

Pricing model
Freemium 1
Audio functions
  • Text-to-speech
  • Voice cloning
  • Voice changer
1
Commercial usage rights
Full commercial use 1
Supported languages
50 languages 1

Pricing

Price from
3 $/mo 1
Free tier
Yes 1

Integration

API available
Yes 1
Report data / suggest a correction

Metrics vs. the category

Supported languages
50 languages Category max 100 languages (12 with data)
Latency (real-time audio)
n/a

Emotionally expressive voice AI: Octave LLM-based text to speech with voice design, cloning and conversion, plus the EVI speech-to-speech interface. Character-based plans with a free tier; Starter from $3/month.

Profile

Hume AI is an emotion-focused voice AI platform built around Octave, a text-to-speech engine that lets you direct the emotional delivery of a line — tone, pacing, emphasis, mood — with plain-language "acting instructions" rather than SSML tags. It is aimed at developers, product teams and creators who need expressive, controllable voice for apps, games, audiobooks and voice agents, alongside a speech-to-speech interface (EVI) and an emotion-measurement API. Pricing runs on a tiered monthly-credit model from a free plan through Enterprise, priced by included characters/minutes rather than a flat seat fee.

Who builds it

Hume AI is a voice and emotion-AI research company founded in 2021 by CEO Alan Cowen, a computational emotion scientist who previously led affective-computing research at Google. The company's core workflow is: pick or design a voice, write a script, optionally attach a natural-language "acting instruction" to steer delivery, and generate speech through Octave (TTS) or hold a live conversation through EVI (speech-to-speech); both sit on top of Hume's emotion-science research.

Core features

  • Octave TTS with acting instructions — direct tone, pacing, emphasis and mood per line using natural-language instructions instead of SSML.
  • Voice cloning and voice design — clone a voice from as little as 15 seconds of sample audio, or design an entirely new voice from a text description.
  • Multilingual generation — Octave 2 (launched October 2025) produces speech in 11 languages (English, Spanish, French, German, Italian, Portuguese, Russian, Arabic, Hindi, Japanese, Korean), versus English and Spanish only on Octave 1.
  • Low-latency streaming — Octave 2 targets roughly 100ms model latency with unidirectional HTTP and bidirectional WebSocket streaming for real-time use.
  • EVI speech-to-speech — a separate empathic voice interface that responds to a caller's tone and prosody in real time, for voice agents and assistants.
  • Word/phoneme-level timestamps and format flexibility — outputs in MP3, WAV or PCM with speed control from 0.25x to 4x.

Pricing

Hume prices by monthly included characters (Octave) and EVI minutes, not by voice count:

  • Free ($0/month): 10,000 TTS characters (~10 minutes), 5 EVI minutes, 15 requests/minute, unlimited voice cloning, non-commercial use only.
  • Starter ($3/month intro): 30,000 characters (~30 minutes), 40 EVI minutes — still non-commercial use only.
  • Creator ($7/month intro, $14/month regular): roughly 140,000 characters and 200 EVI minutes; this is the first tier with a commercial license.
  • Pro ($70/month): 1,000,000 characters (~1,000 minutes), 1,200 EVI minutes, 10 concurrent connections.
  • Scale ($200/month): 3,300,000 characters (~3,300 minutes).
  • Business ($500/month): 10,000,000 characters (~10,000 minutes), 12,500 EVI minutes, 5 seats.
  • Enterprise: custom pricing with unlimited seats, custom rate limits and compliance features.

Paid tiers can buy overage at a fixed rate per 1,000 characters; Free and Starter cannot.

Who it's for

Hume AI fits developers and product teams building voice agents, games or interactive content that need emotionally directed speech rather than flat narration, and researchers or hobbyists who want to test Octave and EVI cheaply. The Free and Starter plans are enough to prototype and evaluate quality but are restricted to non-commercial use with small monthly quotas; anyone shipping a commercial product needs at least the Creator plan, and high-volume production use (long-form audiobooks, large-scale voice agents) points toward Pro, Scale or Business.

Frequently asked questions

What does Hume AI cost?

Hume AI's Octave TTS and EVI speech-to-speech run on a tiered monthly plan priced by included characters and minutes. Free is $0/month (10,000 TTS characters, 5 EVI minutes); Starter is $3/month intro (30,000 characters, 40 EVI minutes); Creator is $7/month intro, $14/month regular (~140,000 characters, 200 EVI minutes) and is the first tier with a commercial license; Pro is $70/month (1,000,000 characters, 1,200 EVI minutes); Scale is $200/month and Business is $500/month, with custom Enterprise pricing above that. Check Hume's current pricing page, since plan limits and intro discounts can change.

Is Hume AI free?

Yes, Hume AI has a genuine $0 Free plan, not just a trial. It includes 10,000 TTS characters (about 10 minutes) and 5 EVI minutes per month, 15 requests/minute, and unlimited voice cloning, but output is restricted to non-commercial use. The Starter plan at $3/month intro raises the quota slightly but is also non-commercial only; a commercial license starts at the Creator plan.

What languages does Hume AI's Octave support?

Octave 2, launched in October 2025, generates speech in 11 languages: English, Spanish, French, German, Italian, Portuguese, Russian, Arabic, Hindi, Japanese and Korean. The earlier Octave 1 model supports only English and Spanish. Hume has said it plans to expand coverage to at least 20 languages, but that had not shipped as of this writing.

Does Hume AI support voice cloning?

Yes. Octave can clone a voice from as little as 15 seconds of sample audio, and separately supports "voice design" — generating a brand-new voice from a natural-language description rather than a recording. Voice cloning is included, unlimited, on every plan including Free, though commercial use of the cloned output requires at least the Creator plan.

Can I use Hume AI's audio commercially?

Only on paid plans from Creator upward. Hume's own TTS FAQ states that Free and Starter subscription tiers are limited to non-commercial use only, while users can use the TTS service for commercial purposes on Creator, Pro, Scale, Business and Enterprise. Note that Hume's terms also grant it a license to use your voice recordings and voice models to provide and improve its services, so it's worth reading the terms of use for specifics before relying on it for client work.

Hume AI vs ElevenLabs — how do they compare?

Both are API-first, expressive TTS platforms with voice cloning, but they differ on entry pricing and control style. Hume's Free plan starts at $0 with 10,000 characters/month and its Starter tier is $3/month intro for 30,000 characters, while ElevenLabs' plans start at $6/month (Starter) with cloning gated to that tier and above rather than on its own free plan. Hume's signature feature is natural-language "acting instructions" for line-by-line emotional delivery, whereas ElevenLabs emphasizes a large pre-built voice library and "Professional Voice Cloning" for higher-fidelity clones on its Creator tier and up. Confirm current limits on each vendor's pricing page before choosing, as both iterate frequently.