# 5 PlayHT alternatives in 2026 for teams whose PlayHT API stopped answering

By Samuel Oram-Jones

A request to PlayHT's API fails before it starts: as of November 2026, a DNS lookup of PlayHT's API host returns no address, a lookup of PlayHT's main site finds nothing either, and the app host presents an expired TLS certificate. PlayHT's documentation site still loads, showing three models on pages marked as updated about a year ago.

That leaves a team with voices, scripts, and a streaming integration and nothing on the other end. Five options follow: Bland Speech, ElevenLabs, Cartesia, Rime, and Deepgram. Bland Speech comes first because Bland Speech v3 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/), and because Bland publishes a migration guide for teams leaving PlayHT. The [text to speech hub](/compare/text-to-speech) and the [alternatives family page](/compare/text-to-speech/alternatives) cover the wider market, and the [LMNT alternatives](/blog/lmnt-alternatives) cover another provider that has stopped serving.

## Why do teams look for PlayHT alternatives?

Teams look for PlayHT alternatives because PlayHT's API host no longer resolves and neither does PlayHT's own pricing site, so a team can't call the API or buy anything.

PlayHT's last documentation shows what a team has to replace. Three models are described: Play 2.0 for legacy projects, Play 3.0 Mini with a published time to first audio of 190ms, and PlayDialog at 350ms, which PlayHT calls its latest voice model for fluid, emotive conversation. PlayHT described an HTTP streaming API and a WebSocket API, input streaming from a language model, instant cloning "with only 30 seconds of speech", Node.js and Python SDKs, and a guide for streaming audio into Twilio calls.

None of it describes something a buyer can use today. A team that depended on PlayHT also needs a way to rebuild its voices, because a cloned voice stays with the vendor that built it and has to be re-cloned from the original recordings, with consent.

## What should I check in a PlayHT alternative?

A buyer should check that the provider is reachable and still shipping, how text streams in, how long a clone sample must be, and what the integration with a phone provider looks like, because those were the four things PlayHT offered.

Reachability is a one-minute test that a team should run before anything else. Streaming differs more. A WebSocket (a connection that stays open so text goes in and audio comes out as both are produced) lets a voice agent send a language model's words as they arrive, and Bland Speech's version lets a new context interrupt the turn that is playing, which is how a caller talking over the agent gets handled.

Clone length is the other number to compare. PlayHT asked for 30 seconds, Bland Speech needs one clean sample of about 10 seconds, and Cartesia and ElevenLabs need 10 seconds as well, on Pro and Starter plans.

Phone audio matters to teams that used PlayHT's Twilio guide, and all five options here return mu-law at 8 kHz, the format phone networks carry. ElevenLabs and Deepgram each publish a Twilio guide, and Cartesia names mu-law for Twilio.

## What are the best PlayHT alternatives?

The best PlayHT alternatives are Bland Speech, ElevenLabs, Cartesia, Rime, and Deepgram, with Bland Speech first for phone calls and voice agents because of its realism result, a WebSocket that takes language model text, and one standard per-character rate.

| Alternative | Best for | Price per million characters | Audio Realism Bench, 4 August 2026 |
| --- | --- | --- | --- |
| Bland Speech | Phone calls and a 10-second instant clone | $40, pay as you go | Bland Speech v3: ranked 1st, 1365 |
| ElevenLabs | A large model range and a Twilio guide | $40 to $80, by model | Eleven v3: ranked 12th, 938 |
| Cartesia | First audio Cartesia puts under 90ms | 1 credit per character, $38 to $65 per million credits past a plan | Sonic 3.5: ranked 6th, 1096 |
| Rime | Pay by the minute, with a $50 monthly free allowance | About $30 for Mist v3 and $50 for Coda v2, from per-minute rates | Wasn't in those results |
| Deepgram | Pay by usage with a $200 free credit | $45 for Flux TTS, $30 for Aura-2 | Wasn't in those results |

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

- **Price per million characters:** each company's published rate with the model or plan named. Rime bills by the minute and says a minute is about 1,000 characters, so its per-character figures are Rime's per-minute rates scaled to a million characters. Deepgram's rates are Pay as you go.
- **Audio Realism Bench:** Design Arena's blind listening test, with rank among 13 AI voices and an Elo score. The scored model is named.

</details>

### 1. Bland Speech

Bland Speech is a text to speech model that Bland runs itself, with a studio to try voices in and an API to build on. Audio streams over HTTP at POST /v2/tts or over a WebSocket at WSS /v2/tts/ws, which takes text as a language model writes it, and a new context immediately preempts the turn that is playing. Bland publishes 199ms to first audio, output is PCM from 8 to 48 kHz or mu-law at 8 kHz, and Pipecat and LiveKit Agents (Python) have integrations for it. Both endpoints are part of Bland's [text to speech API](/text-to-speech-api).

The model suits a team rebuilding a PlayHT voice agent. An instant clone needs one clean sample of about 10 seconds, a professional clone is fine-tuned on 30 minutes or more of audio after Bland's consent and rights review, and every clone is private to the customer's organization. Bland publishes a migration guide that names PlayHT, and [how to migrate from PlayHT](/blog/migrate-from-playht) has the steps.

<div class="cmp-proscons cmp-proscons-own"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>Bland Speech v3 ranked first of 13 AI voices in Audio Realism Bench results of 4 August 2026</li><li>A WebSocket that takes language model text as it is written, with interruption built in</li><li>An instant clone from about 10 seconds of audio, private to the customer's organization</li><li>Mu-law at 8 kHz and PCM from 8 to 48 kHz for phone lines</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 ($40 per million), pay as you go with no monthly plan, and 50,000 free characters once for a new account. Enterprise pricing is custom, set with Bland's sales team. The [Bland Speech](/speech) page has the studio.

### 2. ElevenLabs

ElevenLabs offers HTTP streaming and a WebSocket, a Twilio guide for sending generated audio into a call, and ulaw_8000 output. Eleven v4 and Eleven v4 Turbo launched on 28 September 2026, and the v4 family speaks 90+ languages. The API suits a team that wants a plan ladder: an instant clone from 10 seconds starts on Starter at $6 a month, and a professional clone needs about 30 minutes of audio and the Creator plan at $22.

<div class="cmp-proscons" data-rival="elevenlabs"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>A published Twilio guide and mu-law output at 8 kHz</li><li>Professional clones verified with a voice captcha</li><li>90+ languages across the Eleven v4 family</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</li><li>Eleven v4 bills $0.08 per 1,000 characters, twice Eleven v4 Turbo's $0.04</li><li>The Free plan is non-commercial, and a commercial licence starts on Starter</li></ul></div></div>

**Pricing:** $0.08 per 1,000 characters for Eleven v4 and $0.04 for Eleven v4 Turbo and Flash, billed in dollars on the API. Plans run Starter $6 a month, Creator $22, Pro $99, Scale $299, and Business $990, with Enterprise custom.

### 3. Cartesia

Cartesia sells Sonic 3.6, a model that speaks 44 languages. Cartesia says Sonic delivers first audio in under 90ms, and names pcm_mulaw at 8000 for North American telephony and Twilio. Three endpoints stream audio, bytes over HTTP, SSE, and a WebSocket, and only the WebSocket accepts text that arrives in pieces. The platform suits a team that will manage a monthly credit balance, because Cartesia bills one credit per character from a plan's allowance.

<div class="cmp-proscons" data-rival="cartesia"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>Cartesia says first audio arrives in under 90ms</li><li>Mu-law and A-law output at 8 kHz, with mu-law named for Twilio</li><li>An instant clone from 10 seconds of audio</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Instant cloning needs the Pro plan, and professional cloning needs Startup</li><li>Credits, rollover limits, and overage settings to manage each month</li></ul></div></div>

**Pricing:** Free is $0 a month with 20,000 credits, Pro $5 with 100,000, Startup $49 with 1.25 million, and Scale $299 with 8 million, with Enterprise custom. Overages cost $65, $45, and $38 per million credits on Pro, Startup, and Scale.

### 4. Rime

Rime sells conversational voice models for production voice agents, IVR, and telephony: Coda, released in May 2026, and Mist v3, released in March 2026. Audio streams over HTTP, WebSocket, and gRPC, output includes G.711 mu-law at 8 kHz, and Rime documents integrations with LiveKit, Pipecat, Daily, and Vapi. The platform suits a team that wants to pay by the minute: Rime's Flex plan has no monthly fee, and enterprise customers can run Rime in a dedicated VPC or on their own premises.

<div class="cmp-proscons" data-rival="rime"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>A free plan with $50 of usage every month and no card</li><li>Mu-law output at 8 kHz and word-level timestamps on WebSocket and gRPC</li><li>SOC 2 Type 2 compliant, and HIPAA compliant with a BAA</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Voice cloning is an enterprise order, from 30 to 60 minutes of recordings</li><li>Coda serves 10 cloud API languages, and each voice speaks one</li></ul></div></div>

**Pricing:** Flex is $0.030 a minute for Mist v3 and $0.050 for Coda v2. Scale is $100 a month with 20% off every rate and $100 of credit that rolls over up to $200. Enterprise is custom. Rime publishes two figures for time to first audio: 30ms at P50 for Coda v2 measured at the server without network, and 96ms at P50 for Coda with one request at a time.

### 5. Deepgram

Deepgram's Flux TTS is a voice agent model served on /v2/speak over WebSocket and batch REST, and Deepgram says responses start in as low as 80ms at 24 kHz under production load. Deepgram publishes a guide for Twilio, and says requesting mulaw at 8 kHz "matches Twilio's Media Streams format exactly". The API suits a team that wants to pay by usage: new accounts get a $200 credit with no card and no expiry.

<div class="cmp-proscons" data-rival="deepgram"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>Deepgram says Flux TTS starts speaking in as low as 80ms</li><li>A Twilio guide, and guides for Pipecat and LiveKit</li><li>A $200 free credit with no card and no expiry</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Flux TTS strips SSML markup with a warning and keeps synthesizing</li><li>Flux TTS costs $0.045 per 1,000 characters, 50% more than Aura-2's $0.030</li></ul></div></div>

**Pricing:** Flux TTS is $0.0450 per 1,000 characters on Pay as you go and Growth, and Aura-2 is $0.030 on Pay as you go and $0.027 on Growth. Growth starts at $4,000 a year prepaid, and Enterprise is custom.

## Which PlayHT alternative sounds most human on a call?

Bland Speech v3 ranked first of 13 AI voices at 1365 in Design Arena's Audio Realism Bench results of 4 August 2026, and the human recording scored 1500. The test is blind: a native speaker hears two clips of one line and picks the more human, and Elo, the rating system chess uses, converts the wins and losses into a score.

Cartesia's Sonic 3.5 ranked 6th at 1096 and ElevenLabs' Eleven v3 ranked 12th at 938. PlayHT, Rime, and Deepgram weren't in those results, and Cartesia's Sonic 3.6 (27 August 2026) and ElevenLabs' Eleven v4 (28 September 2026) came after them and weren't scored in them. A blind score says how a voice sounds on a clean recording, so a team should play its own script through 8 kHz mu-law first.

## Which PlayHT alternative should I pick?

Bland Speech is the PlayHT alternative to pick when the goal is a phone agent again: Bland Speech v3 ranked first of 13 AI voices in those 4 August 2026 results, and a clone needs about 10 seconds of audio.

<div class="cmp-usewhen"><div class="cmp-usewhen-own"><p><strong>Use Bland Speech when</strong></p><ul><li>A team will put a cloned or stock voice on phone calls and bill one standard per-character rate with no monthly plan</li><li>A team will stream a language model's reply and let a caller interrupt it</li></ul></div><div><p><strong>Use another option when</strong></p><ul><li>A team will manage plan tiers for a model range and a Twilio guide: ElevenLabs</li><li>A team will manage monthly credit allowances: Cartesia</li><li>A team will pay by the minute and run Rime's models in a dedicated VPC: Rime</li><li>A team will pay by usage from a $200 credit and follow Deepgram's Twilio guide: Deepgram</li></ul></div></div>

On price, Rime's Mist v3 at about $30 per million characters and Deepgram's Aura-2 at $30 per million both sit under Bland Speech's $40, so a team that ranks price over how a caller hears the voice has two cheaper places to start.

## Frequently asked questions

### Is PlayHT still available?

No. PlayHT's API host returns no address in DNS, so a request to it fails, and PlayHT's app host presents an expired TLS certificate. PlayHT's documentation site still loads, with pages marked as updated about a year ago.

### What is the best PlayHT alternative for phone calls?

Bland Speech is the best PlayHT alternative for phone calls, because Bland Speech v3 ranked first of 13 AI voices in Design Arena's results of 4 August 2026, it returns mu-law at 8 kHz, and its WebSocket takes language model text with interruption built in. Bland publishes 199ms to first audio, against PlayHT's published 190ms for Play 3.0 Mini.

### Can a PlayHT cloned voice move to another provider?

No, a cloned voice stays with the vendor that built it, so a team re-clones from the original recordings with consent. PlayHT said a clone needed only 30 seconds of speech, and Bland Speech needs one clean sample of about 10 seconds and the customer's confirmation that the voice is theirs to use.

### Does Bland publish a PlayHT migration guide?

Yes. Bland publishes a migration page that names PlayHT among the providers a team can move from, and [how to migrate from PlayHT](/blog/migrate-from-playht) walks through the steps.

### Which PlayHT alternative costs the least?

Rime's Mist v3 is about $30 per million characters at Rime's rate of 1,000 characters a minute, Deepgram's Aura-2 is $30 per million on Pay as you go, and Bland Speech is $40 per million with 50,000 free characters once. Rime and Deepgram also give free monthly or one-time credit.
