Hume AI
A voice that picks up how the caller is feeling and answers in kind, for warmer and more human conversations.
Paid link, we may earn a commission. How this works.
Scored on the same voice-agent rubric as the full platforms, so a building block like this scores low on the axes it does not address. Read its value score against its job.
See how it stacks up · Full rankings →The expressive one. Hume's EVI generates speech that carries real emotion, and you plug in your own language model behind it. EVI runs about $0.07 a minute to start, less at volume. The catches: you bring your own phone line, and full compliance is Enterprise-only.
About $0.05 to 0.13 for a minute of conversation, once the phone line and the AI are added in.
That's roughly $3.00–7.80 an hour. Plans: from $0/mo (Free) up to $500/mo (Business).
Pricing
Show the cost breakdown
| What the platform charges to run the agent, before the phone line and the AI usage are added on. | $0.07 /min |
|---|---|
| The step that turns what the caller says out loud into text the AI can read. | — |
| The AI 'brain' that reads what the caller said and works out what to say back. | $0.01 /min |
| The step that turns the AI's written reply back into a spoken voice. | — |
| The phone line itself: the service that connects the call to a real phone number. Usually billed on top of the platform. | $0.01 /min |
| The total you actually pay for one minute of conversation once every piece is added up: the platform, the AI, the voice and the phone line. | $0.05–0.13 /min |
EVI (the Empathic Voice Interface) is sold by subscription with included minutes plus a per-minute overage that falls from $0.07/min at the entry tiers to $0.04/min on Business. The language model is bring-your-own (Claude, GPT, Gemini, Llama or a custom endpoint), so its token cost sits on top of the EVI rate; the llm component here is a small BYO estimate, not a Hume charge. Telephony is also bring-your-own Twilio (about $0.014/min), which adds a little latency and drops audio to 8 kHz. Octave text-to-speech is metered separately by characters, $0.15/1k at the entry tiers down to $0.05/1k on Business. HIPAA with a BAA on request is stated on Hume's developer privacy page; SOC 2 Type II and GDPR are listed only as Enterprise features on the pricing page, so they are left unticked here until verified from a certificate.
Every plan in one place: the monthly fee, what each one includes, and the features it unlocks. Anything beyond a plan's allowance, or on a pay-as-you-go tier, is billed at the per-minute rate above. A blank in the features means the vendor's plan page does not state it for that plan, not that it is unavailable.
| Free | Starter | Creator | Pro | Scale | Business | Enterprise | |
|---|---|---|---|---|---|---|---|
| Price | Free | $3/mo | $7/mo | $70/mo | $200/mo | $500/mo | Custom |
| Included | Pay per use | — | — | — | — | — | — |
| Plan notes | 5 EVI min + 10k TTS chars, plus $20 credit; no commercial licence | 40 EVI min included, then $0.07/min | 200 EVI min included, then $0.07/min ($14 regular) | 1,200 EVI min included, then $0.06/min | 5,000 EVI min included, then $0.05/min | 12,500 EVI min included, then $0.04/min | Custom; HIPAA BAA, advertised SOC 2 Type II + GDPR |
| What each plan unlocks | |||||||
| Voice cloning | Yes | Yes | Yes | Yes | Yes | Yes | Yes |
| API access | — | — | — | — | — | — | Yes |
| Concurrent calls | 1 connection | 5 connections | 5 connections | 10 connections | 20 connections | 30 connections | As needed |
| Team seats | — | — | — | — | 3 | 5 | Unlimited |
| Priority support | — | — | — | — | — | — | Slack support |
- Free FreePay per use
5 EVI min + 10k TTS chars, plus $20 credit; no commercial licence
- Voice cloning
- Yes
- API access
- —
- Concurrent calls
- 1 connection
- Team seats
- —
- Priority support
- —
- Starter $3/mo—
40 EVI min included, then $0.07/min
- Voice cloning
- Yes
- API access
- —
- Concurrent calls
- 5 connections
- Team seats
- —
- Priority support
- —
- Creator $7/mo—
200 EVI min included, then $0.07/min ($14 regular)
- Voice cloning
- Yes
- API access
- —
- Concurrent calls
- 5 connections
- Team seats
- —
- Priority support
- —
- Pro $70/mo—
1,200 EVI min included, then $0.06/min
- Voice cloning
- Yes
- API access
- —
- Concurrent calls
- 10 connections
- Team seats
- —
- Priority support
- —
- Scale $200/mo—
5,000 EVI min included, then $0.05/min
- Voice cloning
- Yes
- API access
- —
- Concurrent calls
- 20 connections
- Team seats
- 3
- Priority support
- —
- Business $500/mo—
12,500 EVI min included, then $0.04/min
- Voice cloning
- Yes
- API access
- —
- Concurrent calls
- 30 connections
- Team seats
- 5
- Priority support
- —
- Enterprise Custom—
Custom; HIPAA BAA, advertised SOC 2 Type II + GDPR
- Voice cloning
- Yes
- API access
- Yes
- Concurrent calls
- As needed
- Team seats
- Unlimited
- Priority support
- Slack support
Prices in USD as set by the vendor · last checked 2026-06-03 · vendor pricing →
Slide your expected monthly volume to see roughly what Hume AI would cost.
A rough estimate from Hume AI's sourced rates, not a quote. Always confirm on the vendor's own pricing page before you commit.
At a glance
- Speech-to-text
- Hume (in-house transcription)
- Text-to-speech
- Hume Octave 2 · Bring your own voice: you can upload or clone a custom voice instead of being limited to the platform's stock ones.
- Languages
- en, es, fr, de, it, pt
- Integrations
- Twilio, Anthropic (Claude), OpenAI, Google Gemini, Moonshot Kimi, Llama, Custom LLM endpoint, SDKs (Python / TypeScript / React / .NET)
Compliance
Our full take
Most voice platforms bolt a text-to-speech voice onto a language model and hope the join does not show. Hume takes a different route. Its Octave engine generates the words and the voice together, and it listens for emotion in the caller’s voice, so the agent can sound sympathetic when someone is upset rather than reading every line in the same flat tone. If the feel of the conversation matters to you, this is the one to try first.
The interface is called EVI, the Empathic Voice Interface, and it is sold by subscription with a per-minute rate on top. The plans run from a $3 Starter up to a $500 Business tier, and the more you commit the cheaper each minute gets, from about $0.07 a minute at the entry tiers down to $0.04 at Business. There is a free plan with five minutes and $20 of credit to try it.
Two things are worth understanding before you compare that price to an all-in platform. First, the language model is yours to choose and yours to pay for. Hume lets you bring Claude, GPT, Gemini, Llama or your own endpoint, which is genuinely flexible, but the model’s cost sits on top of the per-minute EVI rate. Second, Hume does not bring a phone line. You connect your own Twilio account, and the phone path adds a little latency and drops the audio quality to standard telephone width.
On voices, Hume’s pitch is breadth through design rather than a fixed library. You can clone a voice from about thirty seconds of audio, or describe a new one in plain words and have it built, and Octave 2 speaks more than sixteen languages. That makes it strong for a branded or character voice, and less obviously so if you just want to pick a stock voice and move on.
Where buyers in regulated industries should slow down is compliance. Hume states clearly that it is HIPAA compliant and will sign a Business Associate Agreement on request, which is real. SOC 2 Type II and GDPR appear only as Enterprise features on the pricing page, with no certificate we can point to, so this page leaves those two unticked until we can verify them. Get the paperwork in writing before you build on anything sensitive.
The 1 to 10 scores here are an editorial preview, our provisional read, not a measured result. We have not call-tested Hume yet, so there is no Voxrater latency figure on this page. The pricing, voices and compliance detail is sourced from Hume’s pricing, product and developer pages, captured 2026-05-31.
Hume AI compared
Our in-depth pieces that put Hume AI side by side with the field, with the sourced numbers and a clear pick.
Alternatives to Hume AI
Other platforms that overlap with Hume AI on the same kind of work, ranked by how many capabilities they share, then by cheaper all-in cost per minute. Compare any of them side by side on the compare page.
Tracking Hume AI? Get the next test result
We re-test and re-price the platforms we cover. Join the list and the next dated update lands in your inbox.
Newsletter launching soon.
Sources
- Hume AI pricing page re-captured 2026-06-02 for the quarterly re-verification; pricing reviewed against the live page (screenshot in evidence/). · captured 2026-06-02
- Hume plan page: per-plan feature matrix (concurrency, seats, voice cloning, API, support) · captured 2026-05-31
- Hume pricing page: EVI tiers $3 to $500/mo, EVI overage $0.07 to $0.04/min, Octave $0.15 to $0.05/1k chars · captured 2026-05-31
- Hume developer privacy page: HIPAA compliant, BAA and DPA on request via legal@hume.ai · captured 2026-05-31
- EVI product page: bring-your-own-LLM list, tool calling, conversational-latency framing · captured 2026-05-31
- EVI 3 announcement: joint voice+language model, voice cloning from ~30s, ~300ms latency claim · captured 2026-05-31
- Octave 2: 16+ languages, voice cloning and text-described voice design · captured 2026-05-31
- Twilio integration: telephony is brought in via Twilio, adds latency and 8 kHz audio · captured 2026-05-31