# 6 best voice cloning APIs in 2026, compared by audio needed and price

By Samuel Oram-Jones

Five of the six voice cloning APIs ranked here ask for about 10 seconds of audio to make an instant clone, and Inworld asks for 3, so sample length sorts them less than the numbers suggest. What separates them is who may make a clone (any account, a $6 monthly plan, or a $49 one), what a clone costs to keep, and whether it can speak on a phone line.

A team looking for a roundup aimed at marketers and creators can read Bland's post on [AI voice cloning platforms for business](/blog/best-ai-voice-cloning-platforms-2026). Developers calling a cloning API from code need a different order, built on audio needed, price per character, consent, and phone-ready output.

## What is the best voice cloning API for developers?

Bland Speech is the best voice cloning API for developers whose clones will speak on phone calls, as of November 2026, because a clone takes one sample of about 10 seconds, a professional clone adds a consent and rights review, a clone's voice ID goes into the same speech calls as any other voice, and those calls return 8 kHz mu-law from a model, Bland Speech v3, that ranked first of 13 AI voices at 1365 in Design Arena's [Audio Realism Bench results of 4 August 2026](https://notes.designarena.ai/audio-realism-benchmark-measuring-realism-in-audio-models/).

That ranking rests on phone output and the realism result, not on audio length or price. Design Arena scored stock voices, one female and one male American-English voice per model, and no clones, which makes its result a score for a model's own voices and not for how closely a clone matches its speaker.

- **Bland Speech:** Best overall for clones that will speak on phone calls
- **ElevenLabs:** Best for self-serve professional cloning with a voice-captcha check
- **Cartesia:** Best for a professional clone with free training
- **Inworld:** Best for cloning from 3 seconds of audio on every plan
- **MiniMax:** Best for paying $1.5 per cloned voice instead of a plan
- **Fish Audio:** Best for API access with no subscription

| API | Audio for an instant clone | Professional clone | Price per million characters | Phone output |
|---|---|---|---|---|
| Bland Speech | About 10 seconds | 30 minutes or more, after a consent and rights review | $40 | 8 kHz mu-law |
| ElevenLabs | 10 seconds on Eleven v4, from the $6 Starter plan | About 30 minutes, from the $22 Creator plan | $80 on v4, $40 on v4 Turbo | 8 kHz mu-law and A-law |
| Cartesia | 10 seconds, up to 60 | From the $49 Startup plan, with free training | $37.38 to $50 in a plan | 8 kHz mu-law and A-law |
| Inworld | 3 seconds, on every plan | At least 10 minutes, in beta | $25 on TTS-2, $15 on TTS-2 Flash | 8 kHz mu-law and A-law |
| MiniMax | At least 10 seconds | One clone type, $1.5 per voice | $100 on speech-2.8-hd, $60 on speech-2.8-turbo | 8 kHz mu-law |
| Fish Audio | At least 10 seconds | At least 10 minutes, in its web app only | $15 per million UTF-8 bytes | mp3, wav, pcm, and opus |

<details><summary>What these rows mean</summary>

- **Audio for an instant clone:** the shortest sample each company documents for a clone made from one recording, with no training run.
- **Professional clone:** a clone trained on a longer recording.
- **Price per million characters:** each company's list price for speech. Cartesia bills 1 credit per character, so its range divides each plan's price by that plan's credits. Fish Audio bills per million UTF-8 bytes, not characters.
- **Phone output:** the formats each company returns that matter on a call. Mu-law at 8 kHz is the format US phone networks carry.

</details>

A developer who needs the shortest sample or the lowest list price should start with Inworld, and a team that wants professional cloning started from a dashboard for $22 a month should look at ElevenLabs first.

## What should a developer check before choosing a voice cloning API?

A developer should check the plan behind the clone before the seconds of audio, because the plan decides who can make one. ElevenLabs puts instant cloning on its Starter plan and professional cloning on Creator, Cartesia offers professional cloning on Startup and above, and Inworld puts instant cloning on every plan, including the one that costs $0 a month. Bland Speech has no monthly plan, and speech is billed per character from a wallet.

Consent comes second, because a clone made from someone who never agreed is a legal problem for the business that ships it. ElevenLabs verifies professional clones with a voice captcha, Fish Audio asks the owner to confirm permission before release, Inworld asks a user to confirm the rights, and Bland asks a customer to confirm every voice and sends professional clones through a consent and rights review. Bland's guide to [consent rules by provider](/blog/is-voice-cloning-legal) covers more.

Phone output is third. A phone call carries 8 kHz mu-law, an 8-bit format that keeps far less of a recording than a studio file does, so a clone should be judged after that conversion. Bland's endpoints are on the [text to speech API](/text-to-speech-api) page.

## How do the 6 best voice cloning APIs compare?

These six APIs differ most in who can clone and what the clone costs to run.

### 1. Bland Speech: Best overall for clones that will speak on phone calls

Bland Speech makes an instant clone from one clean sample of about 10 seconds, in a single file of up to 10 MB, and Bland says it is ready in about a minute. A professional clone is fine-tuned on 30 minutes or more of audio after Bland's consent and rights review. Every clone is private to the customer's organization, and the customer confirms the voice is theirs to use.

A clone's voice ID goes into the same POST /v2/tts and WSS /v2/tts/ws calls as a built-in voice, so a cloned voice streams over a WebSocket that takes language model tokens as they arrive and returns PCM from 8 to 48 kHz or 8 kHz mu-law. Bland Speech suits a team putting a cloned voice on phone calls.

<div class="cmp-proscons cmp-proscons-own"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>An instant clone from one sample of about 10 seconds</li><li>A professional clone from 30 minutes or more, after a consent and rights review</li><li>Clones private to a customer's organization</li><li>8 kHz mu-law and PCM up to 48 kHz</li><li>Bland publishes 199ms to first audio</li><li>Bland Speech v3 ranked first of 13 AI voices on Audio Realism Bench on 4 August 2026</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>More than a hobby project needs, because Bland Speech is built for production calls</li></ul></div></div>

**Pricing:** $0.04 per 1,000 characters, which is $40 per million, pay as you go with no monthly plan, and 50,000 characters free once for a new account. The [Bland Speech](/speech) page has a studio to try a clone.

### 2. ElevenLabs: Best for self-serve professional cloning with a voice-captcha check

ElevenLabs sells instant cloning from its Starter plan at $6 a month, from 10 seconds of audio on Eleven v4, and professional cloning from its Creator plan at $22 a month, from about 30 minutes. For a professional clone the voice owner reads lines from an image aloud and records them, and ElevenLabs verifies that recording. ElevenLabs suits a team that wants clones verified by the voice owner and kept under a monthly plan.

<div class="cmp-proscons" data-rival="elevenlabs"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>Instant clones from 10 seconds of audio on Eleven v4</li><li>Professional clones verified with a voice captcha</li><li>Professional clone fine-tuning for Eleven v4 and v4 Turbo open to every workspace since 2 November 2026</li><li>Mu-law and A-law output at 8 kHz</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Instant cloning needs the $6 Starter plan and professional cloning the $22 Creator plan</li><li>The standard text to speech WebSocket doesn't accept Eleven v3 or v4</li></ul></div></div>

**Pricing:** $0.08 per 1,000 characters for Eleven v4, v3, and Multilingual v2, and $0.04 for v4 Turbo, v3 Conversational, and Flash, against plans from Free to Business at $990 a month. Eleven v3 ranked 12th of 13 AI voices on Audio Realism Bench on 4 August 2026, at 938, and Eleven v4 came out on 28 September 2026, after those results.

### 3. Cartesia: Best for a professional clone with free training

Cartesia sells instant clones from 10 seconds of audio, with up to 60 seconds helping a clone keep an accent on Sonic 3.6. Cartesia offers professional cloning on Startup and above and says training a Pro Voice Clone takes up to 3 hours and costs nothing. [Bland Speech vs Cartesia voice cloning](/compare/bland-speech-vs-cartesia-voice-cloning) compares the two side by side.

<div class="cmp-proscons" data-rival="cartesia"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>Training a Pro Voice Clone is free</li><li>Mu-law and A-law output at 8 kHz</li><li>Cartesia says first audio arrives in under 90ms</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Speed and volume controls don't apply to Pro Voice Clones</li><li>Credits come from a monthly plan, with overages at $65, $45, or $38 per million</li></ul></div></div>

**Pricing:** 1 credit per character from monthly plans: Pro at $5 for 100,000 credits, Startup at $49 for 1.25 million, and Scale at $299 for 8 million.

### 4. Inworld: Best for cloning from 3 seconds of audio on every plan

Inworld clones a voice from as little as 3 seconds of audio, with a 30-second cap, on every plan, including On-Demand at $0 a month. A professional clone, in beta and on TTS-2 only, needs at least 10 minutes of audio and typically takes 30 to 60 minutes to train. Inworld says cloned voices bill at the same rate as stock voices, and Inworld suits a developer who wants the least audio and the lowest per-character price in this list.

<div class="cmp-proscons" data-rival="inworld"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>Instant cloning from as little as 3 seconds on every plan</li><li>Cloned voices bill at the stock-voice rate</li><li>An OpenAI-format speech endpoint, with Pipecat and LiveKit integrations</li><li>Mu-law and A-law output at 8 kHz</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Professional cloning is in beta and runs on TTS-2 only</li><li>Zero data retention doesn't cover clone samples</li></ul></div></div>

**Pricing:** On-Demand is pay as you go at $25 per million characters for TTS-2 and $15 for TTS-2 Flash, with up to 70 minutes of speech included. Monthly plans lower those rates, to $12.50 and $7 per million on Growth at $1,500 a month.

### 5. MiniMax: Best for paying $1.5 per cloned voice instead of a plan

MiniMax clones a voice from a recording of 10 seconds to 5 minutes, up to 20 MB, and charges $1.5 per voice the first time the voice is used for speech. A cloned voice is deleted if it isn't used within 7 days. MiniMax suits a developer who wants to pay per clone and will keep each clone in regular use.

<div class="cmp-proscons" data-rival="minimax"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>$1.5 per cloned voice, charged at first use</li><li>A bidirectional WebSocket that takes language model output as it arrives</li><li>Mu-law output at 8 kHz</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>A cloned voice is deleted after 7 days without use</li><li>Speech costs $60 to $100 per million characters</li></ul></div></div>

**Pricing:** $100 per million characters for speech-2.8-hd and $60 for speech-2.8-turbo on pay as you go, plus $1.5 per cloned voice, or monthly Audio Packages from $5 for 100,000 audio points to $999 for 20 million.

### 6. Fish Audio: Best for API access with no subscription

Fish Audio describes an instant clone from a clip of at least 10 seconds, with a minute or two improving fidelity. A professional clone is made in Fish Audio's web app from at least 10 minutes of clean audio, takes about 1 to 2 hours from start to finish, and needs a paid app plan, with 1 slot on Plus and 5 on Pro. Fish Audio suits a developer who wants a low per-byte price and will convert audio to phone formats on their own side.

<div class="cmp-proscons" data-rival="fish-audio"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>No subscription fees or monthly minimums for API access</li><li>A professional clone in about 1 to 2 hours from start to finish</li><li>s2.1-pro at $15 per million UTF-8 bytes</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Output in mp3, wav, pcm, and opus</li><li>Professional cloning happens in the web app only</li><li>The Free app plan is for personal, non-commercial projects</li></ul></div></div>

**Pricing:** $15 per million UTF-8 bytes for s2.1-pro and $50 for drama-3 on pay as you go.

## How were these voice cloning APIs compared?

Samuel Oram-Jones compared six APIs on seconds of audio for an instant clone, audio and plan for a professional clone, price per million characters, the consent step, and phone-ready output. Rival details come from each company's own pricing, documentation, and terms pages.

Bland's [comparison of text to speech APIs](/compare/text-to-speech) sorts the same companies by job, and its [voice cloning comparisons](/compare/text-to-speech/voice-cloning) take one provider at a time.

## Frequently asked questions

### What is the best voice cloning API?

Bland Speech is the best voice cloning API for clones that will speak on phone calls. A clone takes one sample of about 10 seconds, a professional clone goes through a consent and rights review, and output comes as 8 kHz mu-law. Inworld asks for less audio and charges less per million characters.

### How much audio does a voice cloning API need?

An instant clone needs from 3 seconds at Inworld to about 10 seconds at Bland Speech, ElevenLabs, Cartesia, MiniMax, and Fish Audio. A professional clone needs 30 minutes or more at Bland Speech and about 30 minutes at ElevenLabs, and at least 10 minutes at Inworld and Fish Audio.

### What is the cheapest voice cloning API?

Inworld's TTS-2 Flash at $15 per million characters on pay as you go and Fish Audio's s2.1-pro at $15 per million UTF-8 bytes have the lowest list prices here. MiniMax charges $60 to $100 per million characters plus $1.5 per voice, and Bland Speech charges $40 per million.

### Does a cloned voice cost extra to run?

Pricing differs by company. Inworld bills cloned voices at the stock-voice rate, Cartesia makes professional clone training free and charges 1 credit per character for speech, and MiniMax charges $1.5 per voice at first use. Bland Speech's standard rate is $0.04 per 1,000 characters, in the studio or the API.

### Which voice cloning APIs work on phone calls?

Bland Speech, ElevenLabs, Cartesia, Inworld, and MiniMax return mu-law at 8 kHz, the format US phone networks carry. Fish Audio's API returns mp3, wav, pcm, and opus. Bland's list of the [best voice cloning for phone calls](/blog/best-voice-cloning-for-phone-calls) ranks them for that job.

### Do voice cloning APIs check consent?

Yes, in different ways. ElevenLabs verifies professional clones with a voice captcha, Fish Audio asks the owner to confirm permission before a professional voice is released, and Inworld asks a user to confirm they have the rights. Bland asks a customer to confirm every voice is theirs to use and sends professional clones through a consent and rights review.
