# 5 Typecast alternatives in 2026 for phone calls and voice agents

By Samuel Oram-Jones

Typecast sells a web studio and an API as two separate subscriptions with separate billing, so a team that prototypes in the studio and then builds an app opens a second plan, and the API's Free plan can't be used commercially. On the API, Typecast's Lite plan costs $15 a month for 200,000 credits at one credit per character, as of November 2026.

Bland Speech charges $0.04 per 1,000 characters ($40 per million), pay as you go, with no monthly plan and 50,000 free characters once for a new account, and it ranked first of 13 AI voices in Design Arena's [Audio Realism Bench results of 4 August 2026](https://notes.designarena.ai/audio-realism-benchmark-measuring-realism-in-audio-models/).

## Why do people look for Typecast alternatives?

People look for Typecast alternatives when a studio tool has to become a production API, because Typecast meters the API in credits against a monthly plan, bills it apart from the studio, and keeps cloned voices on its own platform.

Typecast is run by Neosapience in Seoul and describes its API as "built for content automation and app services to conversational agents". Its newest model, ssfm-v31, shipped on 28 October 2026.

The API plans run on credits, one per character. Free gives 15,000 credits a month with attribution required and no commercial use. Lite is $15 a month for 200,000 credits, which Typecast prices at $0.075 per 1,000, and extra use costs $0.09 per 1,000 in blocks of $9. Plus is $280 a month for 4 million credits at $0.07 per 1,000, with extra use at $0.08 in blocks of $100. Those rates work out to $75 and $70 per million characters, against $40 for Bland Speech.

Cloning carries terms a legal team will read. Typecast's API terms let Typecast use submitted voice data to train and improve its models, and a cloned voice can't be exported outside the API. Bland says clones are private to the customer's organization and that the customer confirms every voice is theirs to use.

## What should I check in a Typecast alternative?

A buyer should check the audio format a phone line takes, whether text can stream in while a language model is still writing it, how much audio a clone needs, and what a character costs once the free allowance is gone.

On format, Typecast returns WAV and MP3 at 44,100 Hz and streams WAV at 32,000 Hz, while a phone line carries 8 kHz audio, usually in mu-law, an 8-bit encoding that compresses each sample so a call fits the telephone network. Bland Speech, ElevenLabs, Cartesia, and Inworld each return mu-law at 8 kHz, and Fish Audio returns MP3, WAV, PCM, and Opus.

For streaming, Typecast sends audio back in chunks from a POST endpoint, and a voice agent also needs text to go in a piece at a time as a language model writes it. A WebSocket, a connection that stays open in both directions, allows that, and Bland Speech, ElevenLabs, Cartesia, Inworld, and Fish Audio each offer one.

For cloning, Typecast's API takes an instant clone from 5 to 150 seconds of audio and a professional clone from 5 minutes to 3 hours. For an instant clone, Bland Speech asks for one clean sample of about 10 seconds, Cartesia, ElevenLabs, and Fish Audio ask for 10 seconds or more, and Inworld asks for as little as 3.

## Which Typecast alternative sounds most human?

Bland Speech v3 ranked first of 13 AI voices in Design Arena's Audio Realism Bench results of 4 August 2026, at 1365, and Typecast wasn't in those results, so they put no Typecast voice beside it.

Two other options here had a model in the results. Cartesia's Sonic 3.5 ranked 6th at 1096, and ElevenLabs' Eleven v3 ranked 12th at 938, which is 427 points below Bland Speech v3. Cartesia's Sonic 3.6 came out on 27 August 2026 and ElevenLabs' Eleven v4 on 28 September 2026, so neither was scored. Fish Audio wasn't in those results, and Inworld's one entry was TTS-1.5 Max, ranked 13th at 763, which is not its current TTS-2 line.

Design Arena's listeners hear two recordings of the same text, blind, and pick the one that sounds more like a real person. Each model had one female and one male American English voice, and most prompts ran up to about 40 words, so the results speak to short conversational clips.

## What are the best Typecast alternatives?

The best Typecast alternatives are Bland Speech, ElevenLabs, Cartesia, Inworld, and Fish Audio, with Bland Speech first because it pairs the realism result of 4 August 2026 with phone-ready audio at one per-character rate.

| Alternative | Best for | Published price | Audio Realism Bench, 4 August 2026 |
| --- | --- | --- | --- |
| Bland Speech | Voice agents on phone calls | $40 per million characters, pay as you go | Bland Speech v3: ranked 1st, 1365 |
| ElevenLabs | Audio tags and verified cloning | $40 to $80 per million characters by model, on plans from $6 a month | Eleven v3: ranked 12th, 938 |
| Cartesia | One model across 44 languages | 1 credit per character, on plans from $5 a month | Sonic 3.5: ranked 6th, 1096 |
| Inworld | Plain-words steering at low per-character rates | $15 to $25 per million characters on-demand | TTS-1.5 Max: ranked 13th, 763 |
| Fish Audio | Tag-directed delivery with no subscription | $15 per million UTF-8 bytes on s2.1-pro | Not in those results |

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

- **Best for:** the job each option's published features fit, not a ranking.
- **Published price:** each company's own rate. Bland Speech's $0.04 per 1,000 characters is $40 per million. ElevenLabs charges $0.08 per 1,000 characters on v4 and $0.04 on v4 Turbo and Flash. Cartesia's credits are one per character. Inworld's On-Demand rates are $25 per million for TTS-2 and $15 for TTS-2 Flash. Fish Audio bills per million UTF-8 bytes, and Fish says 1 million bytes is about 180,000 English words.
- **Audio Realism Bench, 4 August 2026:** the model Design Arena scored, its place among the 13 AI voices, and its Elo score, a rating built from blind head-to-head votes. Inworld's entry is an older model than the one it sells now.

</details>

### 1. Bland Speech

Bland Speech is Bland's own text to speech model, v3, sold through a studio and one API and built for phone calls. It streams over HTTP and over a WebSocket that takes text as a language model writes it, and a new context cancels the turn that is playing, so a caller who interrupts stops the voice. Audio comes back as PCM from 8 to 48 kHz or as mu-law at 8 kHz, Bland publishes 199ms to first audio, and an OpenAI-compatible endpoint lets a client built for OpenAI's speech API switch with a new base URL, key, and model. Pipecat has a first-party Bland service, and LiveKit Agents has a Python plugin. More in the [text to speech API](/text-to-speech-api) overview.

Performance tags such as [laughs] and [say warmly] sit before the words they shape, and a pause takes a `<|N|>` marker from 0.1 to 10 seconds. An instant clone takes one clean sample of about 10 seconds, and a professional clone is fine-tuned on 30 minutes or more of audio after Bland's consent and rights review. Bland Speech suits a team building a phone agent or voice agent that wants a realistic voice at a standard rate of $0.04 per 1,000 characters and no monthly plan.

<div class="cmp-proscons cmp-proscons-own"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>Ranked first of 13 AI voices on 4 August 2026, at 1365</li><li>Mu-law at 8 kHz and a WebSocket that takes text as a language model writes it</li><li>$0.04 per 1,000 characters with no monthly plan and 50,000 free characters once</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>More than a narration-only project needs, because Bland Speech is built for production calls</li></ul></div></div>

### 2. ElevenLabs

ElevenLabs sells Eleven v4 and Eleven v4 Turbo, launched on 28 September 2026, alongside older models. ElevenLabs bills the API in US dollars per character, $0.08 per 1,000 on v4 and $0.04 on v4 Turbo and Flash, inside plans that run from a free tier to Business at $990 a month. Audio tags such as [whispering] and [laughing] steer delivery on v4, and the v4 family covers 90+ languages. ElevenLabs says Eleven v4 Turbo's median time to first speech is about 150ms. ElevenLabs suits a team that wants steerable delivery and a clone process with a voice captcha, where the voice's owner reads lines aloud to confirm permission. Instant cloning starts on the Starter plan and professional cloning on Creator.

<div class="cmp-proscons" data-rival="elevenlabs"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>Audio tags for whispering, shouting, and laughing</li><li>90+ languages on the v4 family</li><li>A voice-captcha check for professional clones</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>The standard text to speech WebSocket doesn't take Eleven v3 or v4, so real-time use on v4 goes through Text to Dialogue</li><li>Free users are limited to non-commercial use</li></ul></div></div>

### 3. Cartesia

Cartesia sells Sonic 3.6, a model it says speaks 44 languages, with first audio in under 90ms by Cartesia's own claim. Text goes in over a WebSocket, and Cartesia returns mu-law at 8 kHz for telephony. Plans run on credits at one per character: Pro is $5 a month for 100,000 credits, Startup is $49 for 1.25 million, and Scale is $299 for 8 million, with extra use at $65, $45, and $38 per million credits. Overages are opt-in, and with them off new requests stop when credits run out. Cartesia suits a team that wants one model across many languages and will manage a monthly credit allowance. Instant cloning needs 10 seconds of audio and the Pro plan, and professional cloning needs 30 minutes and the Startup plan.

<div class="cmp-proscons" data-rival="cartesia"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>44 languages on Sonic 3.6</li><li>Sub-90ms first audio, as Cartesia publishes it</li><li>Free training for professional clones</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Credits bill from a monthly plan, and requests stop at zero credits with overages off</li><li>Emotion controls are in beta and English only</li><li>Each break tag costs a credit</li></ul></div></div>

### 4. Inworld

Inworld sells Realtime TTS-2, launched on 5 May 2026, and a faster TTS-2 Flash, launched on 9 August 2026, aimed at consumer-facing apps. On-Demand costs $0 a month and charges $25 per million characters for TTS-2 and $15 for TTS-2 Flash, with up to 70 minutes of speech and a commercial license included. Inworld says an instant clone needs as little as 3 seconds of audio, and both models render 50 non-verbal sounds such as [laugh] and [sigh]. Steering in plain words works on TTS-2 only. Inworld returns mu-law at 8 kHz and suits a team that wants low per-character rates and consumer-style voices.

<div class="cmp-proscons" data-rival="inworld"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>$15 per million characters on TTS-2 Flash, pay as you go</li><li>Instant clones from 3 seconds of audio</li><li>A WebSocket that takes text as a language model writes it</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Steering in plain words works on TTS-2 only, and Flash ignores it</li><li>Zero data retention leaves out voice cloning samples</li></ul></div></div>

### 5. Fish Audio

Fish Audio is a voice platform with a web app, an API, and models whose newest line is Drama 3, launched on 13 October 2026 for voice direction. The production model, s2.1-pro, costs $15 per million UTF-8 bytes with no subscription for API access, and Drama 3 carries a list price of $50 per million bytes with a launch price of $10 through 30 November 2026. Natural-language tags in square brackets, such as [whispers sweetly], steer delivery, and text can stream in over a WebSocket. Fish Audio suits a team directing a performance, for example a character or a narrator, and willing to convert characters to bytes when estimating a bill. A professional clone is made in the web app and needs at least 10 minutes of audio.

<div class="cmp-proscons" data-rival="fish-audio"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>No subscription fee for API access</li><li>Tags that aren't limited to a fixed set</li><li>A free model for development</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Billing per UTF-8 byte means converting characters before budgeting</li><li>Professional clones are made in the web app, not the API</li><li>The app's Free plan is limited to personal, non-commercial use</li></ul></div></div>

## Which Typecast alternative should I pick?

The fact that separates these five is where the audio plays: Typecast's 44,100 Hz WAV and MP3 output suits a studio export or a video, and a call needs audio that leaves the API at 8 kHz and starts before the language model finishes writing.

Bland Speech, ElevenLabs, Cartesia, and Inworld all do that, so the choice among them comes down to the 4 August 2026 result and the billing shape. Bland Speech v3 ranked 1st at 1365 and bills $0.04 per 1,000 characters with no monthly plan.

Pairings are worked through in [Bland Speech vs Typecast](/compare/bland-speech-vs-typecast), [Bland Speech vs ElevenLabs](/compare/bland-speech-vs-elevenlabs), and [Cartesia alternatives](/blog/cartesia-alternatives), and the wider field is in the [text to speech comparison hub](/compare/text-to-speech) and the [alternatives family](/compare/text-to-speech/alternatives). Bland Speech itself is at [bland.ai/speech](/speech).

## Frequently asked questions

### What is the best Typecast alternative?

Bland Speech is the best Typecast alternative for phone calls and voice agents, because Bland Speech v3 ranked first of 13 AI voices in Design Arena's results of 4 August 2026 at 1365 and returns mu-law at 8 kHz at $0.04 per 1,000 characters.

### Is there a cheaper Typecast alternative?

Yes, Inworld's TTS-2 Flash at $15 per million characters and Fish Audio's s2.1-pro at $15 per million UTF-8 bytes cost less per character than Typecast's Lite plan at $75 per million or Bland Speech at $40. Inworld's plain-words steering works on TTS-2 only, and Flash ignores it.

### How much does 1 million characters cost on Typecast?

On the Lite plan, 1 million characters costs $87: the $15 plan covers 200,000 credits, and the remaining 800,000 bill as eight blocks of $9. The Plus plan covers 4 million credits for $280 a month. Bland Speech charges $40 for 1 million characters.

### Does a Typecast alternative need a monthly plan?

Not always. Bland Speech has no monthly plan, Fish Audio charges no subscription fee for API access, and Inworld's On-Demand plan costs $0 a month. ElevenLabs and Cartesia sell their text to speech inside monthly plans.

### Which Typecast alternative clones a voice from the shortest sample?

Inworld says an instant clone needs as little as 3 seconds of audio. Bland Speech takes one clean sample of about 10 seconds, and Cartesia, ElevenLabs, and Fish Audio ask for 10 seconds or more.

### Can a Typecast alternative return phone audio?

Yes, Bland Speech, ElevenLabs, Cartesia, and Inworld each return mu-law at 8 kHz, the encoding most phone lines use. Bland Speech also returns PCM from 8 to 48 kHz.
