# Murf vs WellSaid (2026): Which text to speech platform suits an app, and which suits a voiceover team?

By Samuel Oram-Jones

Murf prices a million characters of Falcon 2.1 speech at $10, billed at $0.01 per 1,000 characters on pay as you go. WellSaid sells differently, scoping each plan with its team "based on how much voiceover you produce", and API access runs from a free account to a trial key to a conversation about the right plan.

That difference sorts the two buyers. Murf suits a team that wants to build speech into an app and know the bill in advance, and WellSaid suits a team producing narration with licensed voices. A team putting speech on a phone line should also weigh Bland Speech, whose v3 model ranked first of 13 AI voices in Design Arena's Audio Realism Bench results of 4 August 2026, results that include neither Murf nor WellSaid.

## Which is better, Murf or WellSaid?

Murf is the better pick for a developer building a real-time voice app, and WellSaid is the better pick for a team producing narration with voices from consenting, paid actors, as of November 2026. Murf publishes $0.01 per 1,000 characters for Falcon 2.1 and says Falcon 2.1 reaches first audio in under 100ms, while WellSaid prices plans with its sales team and positions itself as a voiceover production platform for teams.

**Murf fits when**

- A developer will stream Falcon 2.1 over a WebSocket fed straight from a language model, on pay as you go
- A team will deploy Falcon 2.1 inside its own AWS account from AWS Marketplace, or on its own hardware or VPC with a Murf licence
- A team will pay $19 a month for Murf's Creator voiceover studio with commercial rights, and move to Enterprise to clone a voice

**WellSaid fits when**

- A team will scope a plan with WellSaid's sales team in exchange for voices recorded by actors who gave consent and are paid when their voice is used
- A producer will direct delivery line by line in WellSaid Studio, or with AI Director cues for pitch, tempo, and loudness in the API
- A team will send finished audio to a content management system, digital asset manager, or learning management system through WellSaid Producer, which launched on 10 November 2026

| Feature | Bland Speech | Murf | WellSaid |
|---|---|---|---|
| Audio Realism Bench, 4 August 2026 | Ranked 1st of 13 AI voices, 1365 | Wasn't in those results | Wasn't in those results |
| Price | $0.04 per 1,000 characters ($40 per million), one standard rate | $0.01 per 1,000 characters on Falcon 2.1, $0.03 on Gen2 | Plans scoped with WellSaid's team, API billed by monthly character quota with overage |
| Free to start | 50,000 characters, once | $10 of API credit, no card | A free account and a trial API key |
| First audio, as each company publishes it | 199ms | "Under 100ms" on Falcon 2.1 | "Approximately 500ms per 30 characters" fastest render speed |
| Phone audio | Mu-law at 8 kHz, PCM at 8 to 48 kHz | ALAW and ULAW in HTTP streaming, sample rates from 8,000 Hz | mp3, wav, and s16le, sample rates from 8,000 Hz |
| Instant voice clone | One clean sample of about 10 seconds | Enterprise plan only, one sample up to 30 seconds | Custom voices arranged through WellSaid's team |

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

- **Audio Realism Bench, 4 August 2026:** Design Arena's blind listening test, where listeners hear two clips of the same line and pick the one that sounds more like a real person. Models that weren't entered have no score in it.
- **Price:** WellSaid says it charges "for monthly character usage and overages", which means a quota of characters each month and a bill for anything above it. Murf's per-character rates are on its pay as you go plan.
- **First audio:** The pause between sending text and hearing the first sound. WellSaid's figure is a render speed for text, and WellSaid says streaming reduces the wait.
- **Phone audio:** ALAW and ULAW are the two telephone encodings, A-law for much of the world and mu-law for the US and Japan. WellSaid's s16le is raw 16-bit PCM.
- **Instant voice clone:** Murf's instant cloning is API only and runs on Falcon 2.1, and professional cloning takes up to 90 minutes of studio recordings and 1 to 4 weeks.

</details>

## What are the catches with Murf and WellSaid?

Murf's catches are a single speaking style on Falcon 2.1 and cloning gated to Enterprise. WellSaid's catches are plans that start with a sales conversation, an AI Director that works in English only, and a default API key limited to 1,000 characters per request.

### Murf

Murf sells two text to speech models, Falcon 2.1 for streaming agents and Gen2 for studio-quality narration. Murf decommissioned Gen2 for streaming on 5 October 2026, so Gen2 now serves requests that return a finished file. Falcon 2.1 voices speak in one style, Conversational, and take controls for speed, pitch, and volume plus inline IPA (the International Phonetic Alphabet) and a pronunciation dictionary. Pauses, audio duration, and styles belong to Gen2.

Instant Voice Cloning is "available on the Enterprise plan only", needs one sample of up to 30 seconds, and runs on Falcon 2.1. Professional Voice Cloning is also an Enterprise product, built from up to 90 minutes of studio recordings over 1 to 4 weeks. The Studio Free plan carries no commercial rights.

Murf supports 35 languages across 150-plus voices. Murf says "Murf AI is SOC 2 certified, GDPR compliant, ISO 27001 certified and HIPAA compliant", and describes an active SOC 2 Type II attestation and ISO 27001 certification, with reports under NDA. Enterprise customers get data residency in 11 regional environments.

<div class="cmp-proscons" data-rival="murf"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>$0.01 per 1,000 characters on Falcon 2.1, published, with purchased credit that never expires</li><li>A WebSocket that takes language model tokens directly, with word timestamps</li><li>Official LiveKit and Pipecat integrations</li><li>Data residency and a deployment option in a customer's own AWS account</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Falcon 2.1 speaks in a single style, Conversational</li><li>Instant and professional cloning are Enterprise only</li><li>Gen2 no longer streams</li></ul></div></div>

### WellSaid

WellSaid describes its platform as a way to "Choose a licensed AI voice, direct the final delivery with precision, and produce voiceover content at scale". WellSaid's products are WellSaid Studio, an editor included in every plan, WellSaid Producer, a workflow tool with an MCP server for Claude, Copilot, and ChatGPT, and the WellSaid API. WellSaid says it operates on closed AI models, and that every voice comes from a voice actor who gave consent and is paid when the voice is used.

The API takes a model tag, and the default when none is named is legacy. The next-generation model, caruso, supports AI Director, which steers delivery with pitch, tempo, and loudness tags, and caruso speaks English only. A default API key is limited to 1,000 characters per request. The API has a streaming endpoint that returns audio as it is generated, and WellSaid says streaming "significantly" reduces time to first byte.

WellSaid says it is "SOC2 Type II certified" and that its infrastructure runs on "FINRA, HIPAA and ISO-compliant services". Customer data is processed and stored on Google Cloud in the US, and WellSaid says inputs and outputs are not used to train its model. WellSaid offers voices in 55-plus languages and accents, and partners supply some of the non-English voices.

<div class="cmp-proscons" data-rival="wellsaid"><div class="cmp-proscons-pros"><p><strong>Pros</strong></p><ul><li>Every voice recorded by a consenting, paid voice actor</li><li>Line-by-line direction in Studio, and AI Director cues in the API</li><li>Producer sends finished audio to a CMS, DAM, or LMS</li><li>SOC 2 Type II and a no-training statement</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>Plans are scoped with WellSaid's team, and API access ends in a plan conversation</li><li>AI Director runs on the caruso model, English only</li><li>A default API key allows 1,000 characters per request</li><li>Custom voices are arranged through WellSaid's team</li></ul></div></div>

## What do Murf, WellSaid, and Bland Speech cost?

Murf charges $0.01 per 1,000 characters for Falcon 2.1 and $0.03 for Gen2, WellSaid scopes plans with its team, and Bland Speech charges $0.04 per 1,000 characters ($40 per million) with no monthly plan.

### Murf pricing

Murf's API starts with a $10 free credit and no card. Pay as you go bills Falcon 2.1 at $0.01 per 1,000 characters and Gen2 at $0.03, and Murf says "Purchased credit never expires". Speech to text in beta costs $0.12 an hour, and Murf offers early-stage startups 50 million free characters over 3 months. Enterprise API pricing is custom, with an SLA, volume discounts, and on-premise deployment.

Murf's Studio, the voiceover app, has a Free plan of 10 minutes with no commercial rights, Creator at $19 a month ($228 billed annually), Business at $66 a month ($792 billed annually), and a custom Enterprise plan.

### WellSaid pricing

WellSaid says Studio "is part of every WellSaid plan" and that plans are scoped by how much voiceover a team produces. API access begins with a free account and a trial key, and after the trial WellSaid says to "work with our team to find the right plan for your business". API billing is a monthly character quota with overage charges.

### Bland Speech pricing

Bland Speech costs $0.04 per 1,000 characters, pay as you go, with the same standard rate in the studio and the API. A new account gets 50,000 free characters once, and credits don't expire. Each request carries a minimum charge of $0.001, and some public-library voices add a per-character creator fee. Murf's rates run below Bland Speech's, and none of the three rates includes a phone line or taxes.

### What a month costs

A million characters of speech costs $10 on Murf's Falcon 2.1, $30 on Murf's Gen2, and $40 on Bland Speech, and a WellSaid price depends on the plan WellSaid scopes.

## Which sounds more human, Murf, WellSaid, or Bland Speech?

In Design Arena's Audio Realism Bench results of 4 August 2026, Bland Speech v3 ranked first of 13 AI voices at 1365 against a human recording at 1500, and neither Murf nor WellSaid was in those results.

Design Arena's listeners compare two unlabelled clips of the same line and pick the more human one, and the scores are Elo ratings, the chess-style system in which beating a strong opponent earns more than beating a weak one. Phone-agent prompts are one category in the test. A voice that wasn't entered has no score, so the results say nothing about how Murf or WellSaid sounds, and Murf's Falcon 2.1 launched on 19 October 2026, after them. Design Arena explains its method in [its Audio Realism Bench article](https://notes.designarena.ai/audio-realism-benchmark-measuring-realism-in-audio-models/).

## Why does Bland Speech suit phone calls better than Murf or WellSaid?

Bland Speech suits phone calls better than Murf or WellSaid because Bland built it for phone calls, and Bland Speech v3 ranked first in the 4 August 2026 results. Murf's published rate is lower than Bland Speech's $40 per million characters, and Murf says Falcon 2.1 reaches first audio in under 100ms, faster than the 199ms Bland publishes.

Bland Speech's WebSocket at /v2/tts/ws takes text as a language model writes it and drops the turn that is playing when a new context arrives, so a caller can interrupt. Output is mu-law at 8 kHz or PCM from 8 to 48 kHz, and a first-party Pipecat service and a LiveKit Agents plugin for Python cover the two frameworks Murf also integrates with. An endpoint that follows OpenAI's speech contract lets a team switch by changing a base URL, a key, and a model name. The [text to speech API](/text-to-speech-api) page has the endpoints.

Murf says it is SOC 2 certified and HIPAA compliant, and WellSaid says it is SOC 2 Type II certified. 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, and Bland runs its own speech model. A clone on Bland Speech starts from one clean sample of about 10 seconds, and the customer confirms the voice is theirs to use.

<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 Design Arena's Audio Realism Bench results of 4 August 2026</li><li>One standard rate of $0.04 per 1,000 characters, with no plan to scope and no monthly minimum</li><li>Mu-law and PCM for phone calls, with interruption built into the WebSocket</li><li>An OpenAI-compatible endpoint, a Pipecat service, and a LiveKit Agents plugin</li></ul></div><div class="cmp-proscons-cons"><p><strong>Cons</strong></p><ul><li>More than a narration project needs, because Bland Speech is built for production calls</li></ul></div></div>

More pairings sit on the [rival comparisons](/compare/text-to-speech/rival-vs-rival) page, and the [text to speech comparison hub](/compare/text-to-speech) covers every vendor. Bland has pages on [Bland Speech against Murf](/compare/bland-speech-vs-murf) and [Bland Speech against WellSaid](/compare/bland-speech-vs-wellsaid), plus the [Murf AI review](/blog/murf-ai-review) and the [WellSaid review](/blog/wellsaid-review). The [Bland Speech](/speech) page has the studio and the price.

## Frequently asked questions

### Is Murf better than WellSaid?

Murf is better for a developer who wants a published per-character rate and a streaming API, and WellSaid is better for a team that wants voices from consenting, paid actors and a production workflow. Neither was in Design Arena's 4 August 2026 results, and a phone agent also has Bland Speech to weigh.

### Which is cheaper, Murf or WellSaid?

Murf prints its prices and WellSaid scopes plans with its team, so only Murf's can be compared in advance. Murf's Falcon 2.1 costs $10 per million characters on pay as you go and Gen2 costs $30, against $40 for Bland Speech.

### Can Murf and WellSaid clone a voice?

Murf's instant clone is an Enterprise feature that takes one sample of up to 30 seconds, and Murf's professional clone takes up to 90 minutes of recordings over 1 to 4 weeks. WellSaid builds custom digital voice replicas for a brand through its team, with consent and licensing on record. Bland Speech's instant clone needs one clean sample of about 10 seconds.

### How do Murf and WellSaid stream audio?

Murf offers HTTP streaming and a WebSocket on Falcon 2.1, and the WebSocket takes language model tokens directly. WellSaid's API has a streaming endpoint that returns audio as it is generated.

### Do Murf and WellSaid return phone-ready audio?

Murf's HTTP streaming returns ALAW and ULAW among its formats, with sample rates from 8,000 Hz. WellSaid's streaming endpoint returns mp3, wav, or raw PCM as s16le, with sample rates from 8,000 Hz. Bland Speech returns mu-law at 8 kHz or PCM at 8 to 48 kHz.

### What are the security credentials of Murf and WellSaid?

Murf says it is SOC 2 certified, GDPR compliant, ISO 27001 certified, and HIPAA compliant, and WellSaid says it is SOC 2 Type II certified and complies with GDPR and the EU AI Act. Bland is SOC 2 Type II audited and HIPAA compliant with a signed BAA.

### Which text to speech platform suits a voice agent on a phone line?

Bland Speech suits a voice agent on a phone line better than either, because Bland Speech v3 ranked first of 13 AI voices on 4 August 2026, returns mu-law at 8 kHz, and lets a new context interrupt the turn that is playing. Murf's Falcon 2.1 is the closer match among the two rivals for a developer who wants streaming.
