# Bland Speech vs Resemble AI (2026): Does Resemble AI still sell text to speech?

A team that searches for Resemble AI to buy text to speech runs into a problem before any comparison starts: Resemble AI has stopped selling it. In a post dated 13 August 2026, its co-founder and CEO, Zohaib Ahmed, wrote "We're not selling voice AI to new customers", and the company now sells deepfake detection and watermarking.

That turns this into a different comparison. Bland Speech is a text to speech API a team can sign up for today, with 50,000 free characters. Resemble AI's remaining voice offering is open-source model code, which Resemble AI calls retired.

## Which is better, Bland Speech or Resemble AI?

For text to speech, Bland Speech is the one a new customer can buy as of November 2026. Resemble AI says its generative voice models "are no longer offered as commercial paid products", so there is nothing from Resemble AI to compare on price, streaming, or phone audio. Bland Speech v3 ranked first of 13 AI voices in Design Arena's Audio Realism Bench results of 4 August 2026, and Resemble AI wasn't in those results.

Resemble AI is the better choice for a different job: detecting AI-generated audio, video, and images, or enrolling a voice for identity matching. Bland's [comparison of text to speech APIs](/compare/text-to-speech) and its [head-to-heads](/compare/text-to-speech/head-to-heads) cover vendors that do sell text to speech.

<div class="cmp-usewhen"><div class="cmp-usewhen-own"><p><strong>Use Bland Speech when</strong></p><ul><li>A team needs a text to speech API it can start using today, with 50,000 free characters</li><li>Callers on a phone line should hear a voice that sounds like a person</li><li>A team is moving off Resemble AI's retired voice products and will re-create its voices from original recordings</li><li>A team wants a private voice clone from one 10-second sample, with consent confirmed for every voice</li></ul></div><div><p><strong>Use Resemble AI when</strong></p><ul><li>A team will detect AI-generated audio, video, and images in real time as part of a call or meeting workflow</li><li>A team will enroll a voice or likeness for identity matching</li><li>A team will host Chatterbox itself under the MIT licence and carry its own support</li><li>A team will scope an Enterprise detection contract with on-premise deployment</li></ul></div></div>

| Feature | Bland Speech | Resemble AI |
|---|---|---|
| Text to speech for new customers | Yes, as an API and a studio | "Not selling voice AI to new customers" |
| Audio Realism Bench, 4 August 2026 | Ranked 1st of 13 AI voices, 1365 | Not in those results |
| Price | $0.04 per 1,000 characters ($40 per million) | Detection plans: Flex $0, Team $350 a month, Business $1,000 a month, Enterprise custom |
| Free usage | 50,000 characters, once | Flex plan at $0, for detection |
| Phone audio | mu-law at 8 kHz, and PCM from 8 to 48 kHz | Not sold to new customers |
| Voice cloning | Instant from about 10 seconds, professional from 30 minutes or more | Not sold to new customers |
| Open source | No: a closed-source model Bland built itself | Chatterbox, MIT licence, retired in August 2026 |

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

- **Text to speech for new customers:** Resemble AI wrote "We're supporting our existing customers but not taking new ones." It also says its generative voice models are "no longer offered as commercial paid products".
- **Audio Realism Bench, 4 August 2026:** Design Arena's blind listening test, where listeners hear two clips of the same line from two models and pick the one that sounds more like a real person. A higher score means listeners picked that voice more often.
- **Price:** Bland Speech costs $0.04 per 1,000 characters, which is $40 per million. Resemble AI's processing rates are for detection, identity, watermarking, and agent detection.
- **Phone audio and voice cloning:** Resemble AI says it does not sell text to speech, voice cloning, or any other voice AI product. Bland Speech returns mu-law, the 8-bit telephony format, at 8 kHz.
- **Open source:** Chatterbox is a family of open-source text to speech models by Resemble AI. Resemble AI says Chatterbox has been retired research since August 2026, with no support, updates, or paid tier.

</details>

## How do Bland Speech and Resemble AI compare?

The biggest difference is what each company sells now. Bland sells a text to speech model built for phone calls, where polish reads as fake. Resemble AI sells detection, and says it detects AI-generated audio, video, and images in real time with explainable results.

### Bland Speech overview

Bland Speech is the text to speech model inside Bland, a voice AI provider that runs its own model and speech rather than wrapping third parties, and it is sold on its own as a studio and an API on the [Bland Speech](/speech) page.

A team sends text over HTTP to POST /v2/tts and gets raw audio or a WAV file back, or streams language model output into a WebSocket at WSS /v2/tts/ws, where a new context interrupts the turn that is playing. POST /v2/audio/speech follows the OpenAI audio speech contract, and the [text to speech API](/text-to-speech-api) page has the code. Performance tags in square brackets, such as [laughs], [sighs], or [say warmly], tell the model how to deliver the words after them.

<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 on Audio Realism Bench on 4 August 2026</li><li>One standard rate, $0.04 per 1,000 characters, with no monthly plan</li><li>Open to new customers, with 50,000 free characters</li><li>Mu-law at 8 kHz and PCM from 8 to 48 kHz, ready for phone calls</li><li>Instant clones from about 10 seconds of audio, private to the customer's organization</li><li>A first-party Pipecat service and a Python plugin for LiveKit Agents</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>

### Resemble AI overview

Resemble AI is still trading, with its attention on deepfake detection: Resemble AI says it returns a verdict on audio, video, and image in under 300 milliseconds. It supports existing voice customers, has retired more than 300 older posts on voice cloning, text to speech, and audio editing, and keeps its old voice API guides as history only. Its open voice models, Chatterbox among them, stay free to use, and Resemble AI says they have been downloaded more than 17 million times on Hugging Face.

Chatterbox is the open-source family: Chatterbox Multilingual V3 with 23 languages, Chatterbox-Turbo for low-latency English agents, and Chatterbox-Nano for CPUs. Every file Chatterbox generates carries Resemble AI's PerTh watermark.

<div class="cmp-proscons" data-rival="resemble-ai"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>Deepfake detection and watermarking for audio, video, and images, with a verdict in under 300 milliseconds</li><li>Chatterbox weights are free under the MIT licence</li><li>Every Chatterbox output carries a PerTh watermark</li><li>Voice and likeness enrolment for identity matching, and SOC 2 documentation on Enterprise</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Not selling voice AI to new customers, as of 13 August 2026</li><li>Chatterbox is retired research with no support, updates, or paid tier</li><li>Voice content retired from the site, including more than 300 older posts</li></ul></div></div>

### What should a team using Resemble AI voices check?

Resemble AI says it is supporting existing customers, and it keeps its old voice API guides as history only, saying they "do not imply continued availability of commercial voice services." A team already on Resemble AI's hosted voices has support today and no written promise beyond it, and a team starting a new text to speech project has no Resemble AI product to buy.

Moving a voice is the hard part. A cloned voice made on one provider's model doesn't carry over to another provider, so a team re-creates each voice from its original recordings, with consent for every one. Bland publishes a migration guide for moving from Resemble AI to Bland Speech, and Bland Speech's instant clone needs one clean sample of about 10 seconds, ready in about a minute. A professional clone is fine-tuned on 30 minutes or more of audio after a consent and rights review, and every clone is private to the customer's organization.

### How do Bland Speech and Resemble AI handle security?

Bland says on its [trust page](/trust-security) that speech synthesis runs inside its FedRAMP boundary. Bland is SOC 2 Type II audited, HIPAA compliant with a signed BAA, and FedRAMP 20x Class C certified at Moderate, granted on 17 November 2026. Enterprise customers get dedicated orchestration and dedicated GPUs run by Bland.

Resemble AI's pricing for detection names SOC 2 documentation as an Enterprise item, alongside on-premise deployment and volume pricing. Open Chatterbox weights run on a team's own infrastructure, so no vendor security terms apply to them.

## How much do Bland Speech and Resemble AI cost?

Bland Speech costs $40 per million characters, and no text to speech price applies to Resemble AI because it sells none to new customers. Chatterbox is free to download under the MIT licence, and the compute to run it is the team's own.

### Bland Speech pricing

Bland Speech costs $0.04 per 1,000 characters, directed or not, in the studio or the API. New accounts get 50,000 characters free, and credits do not expire. A 60-second note costs about $0.04, a 10-minute narration about $0.36, and a million characters $40.

Enterprise pricing is custom, set with Bland's sales team.

### Resemble AI pricing

Resemble AI's pricing covers detection and verification: Flex at $0, Team at $350 a month, Business at $1,000 a month, and Enterprise with custom volume pricing, enterprise SLAs, and on-premise deployment. Chatterbox has no hosted price, and Resemble AI keeps the code public with no paid tier.

A team weighing self-hosted Chatterbox against a $40 per million characters API is weighing GPU hours and engineering time against a bill that scales with text. With no support or updates behind the model, that engineering time stays with the team.

## Which sounds more human, Bland Speech or Resemble AI?

Bland Speech v3 scored 1365 in the blind tests Design Arena published on 4 August 2026, first of 13 AI voices, with human recordings at 1500. Resemble AI wasn't in those results, and its hosted voices are closed to new customers, so a new team has no Resemble AI service to audition.

Listeners in Audio Realism Bench hear two clips of the same line, without knowing which model made which, and pick the one that sounds more like a real person. In Design Arena's words, "Second only to human speech is Bland Speech v3 by Bland AI, with an Elo of 1365", and the full chart is in [Design Arena's Audio Realism Bench article](https://notes.designarena.ai/audio-realism-benchmark-measuring-realism-in-audio-models/).

## Why is Bland Speech the best Resemble AI alternative?

A team leaving Resemble AI needs text to speech it can start on today, and Bland Speech is that, from a company that runs its own model and speech, with a voice that ranked first of 13 AI voices on Audio Realism Bench on 4 August 2026. Bland Speech returns mu-law at 8 kHz for phone lines, bills one standard rate with no monthly plan, and takes an OpenAI-style request after a change of base URL, key, and model name.

Pipecat maintains a Bland text to speech service and LiveKit Agents has a Bland Python plugin, so a voice agent built on either framework swaps providers inside the framework it already runs. [Bland Speech vs Resemble AI voice cloning](/compare/bland-speech-vs-resemble-ai-voice-cloning), the [Resemble AI alternatives](/blog/resemble-ai-alternatives), the [Resemble AI review](/blog/resemble-ai-review), and [Bland Speech vs Chatterbox](/compare/bland-speech-vs-chatterbox) go further.

### How was this comparison made?

The criteria were what each company sells, voice realism, price, voice cloning, open-source models, and security. Resemble AI's facts come from its own site, pricing, documentation, and Chatterbox repository, and from its CEO's post of 13 August 2026, and voice scores come from Design Arena's published results of 4 August 2026.

## Frequently asked questions

### Does Resemble AI still sell text to speech?

No. In a post dated 13 August 2026, Resemble AI's CEO wrote "We're not selling voice AI to new customers", and Resemble AI says it does not sell text to speech, voice cloning, or any other voice AI product. It supports existing customers and sells deepfake detection.

### Is Bland Speech better than Resemble AI?

For text to speech, yes, because a new customer can buy Bland Speech and can't buy it from Resemble AI. Bland Speech v3 ranked first of 13 AI voices on Audio Realism Bench on 4 August 2026, and Resemble AI wasn't in those results. Resemble AI suits teams that need deepfake detection.

### Is Chatterbox still maintained?

No. Resemble AI says Chatterbox has been retired research since August 2026 and is no longer maintained, with the code kept public under the MIT licence and no support, updates, or paid tier.

### What replaces Resemble AI voice cloning?

A team re-creates each voice from original recordings on a provider that sells cloning. Bland Speech makes an instant clone from one clean sample of about 10 seconds and a professional clone from 30 minutes or more of audio, after a consent and rights review, and every clone is private to the customer's organization.

### Does Bland Speech work on phone calls?

Yes. Bland built Bland Speech for phone calls, and it returns mu-law at 8 kHz as well as PCM from 8 to 48 kHz, so audio goes onto a phone line without conversion.
