# Amazon Polly — Pay-per-character cloud text to speech from AWS, with four voice engines and a browser console.

> Source: Dubsheet — https://dubsheet.com/amazon-polly (structured, researched, re-verified)

Amazon Polly is AWS's cloud text-to-speech service that turns written scripts into downloadable speech through an API, a command line or a browser console. It offers four voice engines, dozens of named voices and markup tags that shape pronunciation and delivery.

**Best for:** Developers adding spoken output to apps, e-learning and phone systems

| Fact | Value |
| --- | --- |
| Website | https://aws.amazon.com/polly/getting-started |
| Category | Text to speech APIs, Cheapest for 2 hours a month, Courses and e-learning, Free text to speech |
| API | Yes |
| Open source | No |
| Price | Free |
| Price checked | 27 October 2026 |

Screenshot of Amazon Polly: https://dubsheet.com/img/1920/screenshots/amazon-polly.webp, captured 28 October 2026

**What we checked**

- Pages last read: 27 October 2026
- Price on its site: Free plan (From its pricing page)

## Amazon Polly price and terms

**Terms**

- Free plan: Yes — 5 million characters a month on Standard voices — 1 million Neural, 500,000 Long-Form and 100,000 Generative characters a month for the first 12 months. New accounts since 15 July 2025 also get up to $200 in AWS Free Tier credits, usable within 12 months of sign-up.
- Sign up without sales: Yes

## For AI agents

- API: Yes

Pricing page as we captured it: https://dubsheet.com/img/1920/screenshots/amazon-polly/pricing.webp, captured 28 October 2026

Read on https://aws.amazon.com/polly/pricing/, 27 October 2026.

## What we checked on Amazon Polly

### What an hour of speech costs

**Cost for 2 hours a month** · Yes — $0 a month on AWS Free Tier - Standard voices for 2 hours of 60 minutes. credits counted per character or per word are converted at the standard hour: 52,000 characters, 9,300 words

> For Amazon Polly's Standard voices, the free tier includes 5 million characters per month for speech or Speech Marks requests.
> — https://aws.amazon.com/polly/pricing/ · Free tier section, 'Millions of characters per month', checked 27 October 2026

**Cost for 10 hours a month (heavy use)** · Yes — $0 a month on AWS Free Tier - Standard voices for 10 hours of 60 minutes. credits counted per character or per word are converted at the standard hour: 52,000 characters, 9,300 words

> For Amazon Polly's Standard voices, the free tier includes 5 million characters per month for speech or Speech Marks requests.
> — https://aws.amazon.com/polly/pricing/ · Free tier section, 'Millions of characters per month', checked 27 October 2026

**Promotion and regular price** · No — none — the published rate is the standing pay-as-you-go price per million characters, billed monthly for what you used, with no introductory discount and no annual commitment shown on the pricing page.

> Pay-as-you-go model You are billed monthly for the number of characters of text that you processed.
> — https://aws.amazon.com/polly/pricing/ · Pricing section, 'Pay-as-you-go model', checked 27 October 2026

**What the price counts** · Yes — Billing counts characters of input text, not minutes or credits: $4.00 per million characters for Standard voices, $16.00 Neural, $30.00 Generative and $100.00 Long-Form, with no monthly plan allowance to buy.

> Amazon Polly's Standard voices are priced at $4.00 per 1 million characters for speech or Speech Marks requests (when outside the free tier).
> — https://aws.amazon.com/polly/pricing/ · Pricing section, 'Pay-as-you-go model' paragraph, checked 27 October 2026

**Voices that use more** · Yes — The newer and higher-quality voices cost far more per character: Neural is four times the Standard rate, Generative seven and a half times, and Long-Form twenty-five times, so voice choice drives the bill.

> Amazon Polly's Long-Form voices are priced at $100.00 per 1 million characters for speech or Speech Marks requested (when outside the free tier).
> — https://aws.amazon.com/polly/pricing/ · Pricing section, 'Pay-as-you-go model' paragraph, checked 27 October 2026

**Past your allowance** · Yes — There is no plan to exhaust: once the free tier allowance for a voice type is used up, every further character is billed at that voice type's published per-million rate, and requesting timing data is billed again.

> Yes. You will be charged for every request for speech or Speech Marks based on the number of characters you send to the service.
> — https://aws.amazon.com/polly/faqs/ · Billing section, question about 1,000 characters plus Speech Marks, checked 27 October 2026

**Commercial use** · Yes — Output may be published and resold from the first character, free tier included: the vendor states no restrictions on storing or reusing generated speech and confirms the audio belongs to you.

> Amazon Polly's pay-as-you-go pricing, low cost per request, and lack of restrictions on storage and reuse of voice output make it a cost-effective way to enable speech synthesis everywhere.
> — https://aws.amazon.com/polly/faqs/ · General section, 'What is Amazon Polly?', checked 27 October 2026

### The voices

**Stock voices** · Yes — The vendor describes dozens of lifelike voices across four engines, and its own table names them one by one with a listenable sample for each voice and variant.

> Amazon Polly includes dozens of lifelike voices and support for a variety of languages, so you can select the ideal voice and distribute your speech-enabled applications in many countries.
> — https://aws.amazon.com/polly/features/ · 'Wide Selection of Voices and Languages' section, checked 27 October 2026

**Languages** · Yes — Its voice table covers roughly thirty languages with named regional accents, including US, UK, Australian, Indian, Irish, New Zealand, Singapore and South African English, plus Arabic, Cantonese, Mandarin, Japanese and Korean.

> Amazon Polly includes dozens of lifelike voices and support for a variety of languages, so you can select the ideal voice and distribute your speech-enabled applications in many countries.
> — https://aws.amazon.com/polly/features/ · 'Wide Selection of Voices and Languages' voice table, Language or Language variant column, checked 27 October 2026

**Re-voice your recording** · No — It converts written text only; you send text and get audio back, so there is no way to record your own line and have it returned in another voice with your timing.

> Amazon Polly is easy to use - you just send the text you want converted into speech to the Amazon Polly API, and Amazon Polly immediately returns the audio stream to your application
> — https://aws.amazon.com/polly/faqs/ · General section, 'What is Amazon Polly?', checked 27 October 2026

**Emotion, pace and emphasis** · Yes — Markup tags set speaking rate, pitch, volume, pauses, emphasis and a whisper effect, and some Neural voices can be switched to a newscaster reading style.

> You can control various aspects of speech such as pronunciation, volume, pitch, speech rate, etc. using standardized Speech Synthesis Markup Language (SSML).
> — https://aws.amazon.com/polly/faqs/ · General section, 'What features are available?', checked 27 October 2026

**Pronunciation fixes** · Yes — Custom pronunciation dictionaries fix names, acronyms and foreign words once and are reapplied to every request, with up to 100 stored per account and five applied per text.

> With Amazon Polly's custom lexicons, or vocabularies, you can modify the pronunciation of particular words, such as company names, acronyms, foreign words and neologisms
> — https://aws.amazon.com/polly/features/ · 'Custom Lexicons' section, checked 27 October 2026

**Long scripts and chapters** · Partly — Long scripts go through the asynchronous task operation, capped at 100,000 billed characters per submission, while a single direct request takes 3,000 billed characters and ten minutes of audio; there are no chapters or per-line regeneration.

> The size of the input text can be up to 100,000 billed characters (200,000 total characters).
> — https://docs.aws.amazon.com/polly/latest/dg/limits.html · 'SpeechSynthesisTask API operations' section, checked 27 October 2026

**Several speakers** · No — One request speaks in one voice: the markup tag that would switch voices inside a document is listed as unsupported, so dialogue or a two-host show needs separate requests stitched together yourself.

> The `<audio>`, `<lexicon>`, `<lookup>`, and `<voice>` tags are not supported.
> — https://docs.aws.amazon.com/polly/latest/dg/limits.html · Speech Synthesis Markup Language limits section, checked 27 October 2026

**Audio you get** · Yes — Downloads and streams come as MP3, Ogg Vorbis or raw PCM, with a choice of sampling rates up to 24 kHz so you can trade file size against audio quality.

> Amazon Polly supports MP3, Vorbis, and raw PCM audio stream formats.
> — https://aws.amazon.com/polly/features/ · 'Optimize Your Streaming Audio' section, checked 27 October 2026

### Cloning a voice

**Instant cloning** · Partly — A custom voice exists only as Brand Voice, a bespoke engagement where the vendor's team records a chosen actor and trains the model; there is no self-serve clone from a short sample you upload.

> Brand Voice is a custom engagement where you work with the Amazon Polly team to build an Neural Text-to-Speech (NTTS) voice for the exclusive use of your organization.
> — https://aws.amazon.com/polly/features/ · 'Brand Voice' section, checked 27 October 2026

**Professional cloning** · Partly — Brand Voice is the high-fidelity option: the vendor records an actor, trains a model and releases the voice to your account alone, but it is quoted case by case rather than sold on the price list.

> We work with you throughout the entire process to identify the persona, identify an actor or actress and record their speech, and ultimately build and train a model to produce the voice.
> — https://aws.amazon.com/polly/features/ · 'Brand Voice' section, checked 27 October 2026

**Your voice trains their models** · Yes — The vendor says it may keep and use the text you send to improve its own machine learning, and an account-level opt-out policy is documented for turning that off.

> Amazon Polly may store and use text inputs processed by the service solely to provide and maintain the service and to improve and develop the quality of Amazon Polly and other Amazon machine-learning/artificial-intelligence technologies.
> — https://aws.amazon.com/polly/faqs/ · Data Privacy section, first question, checked 27 October 2026

### Dubbing and translation

**Video dubbing** · No — It never takes a video: the vendor's own localization example has you translate the text with a separate service and then voice it here, so dubbing is something you assemble around the product.

> Let's say you translate the text using Amazon Translate and voice it with Polly.
> — https://aws.amazon.com/polly/features/ · 'Adjust the Maximum Duration of Speech' section, checked 27 October 2026

**Lip sync** · Partly — It does not touch video, but it returns lip-shape and word timing data alongside the audio so your own app can animate an avatar's mouth in step with the speech.

> Viseme, which describes the shape of the lips that corresponds to the sound that is spoken
> — https://aws.amazon.com/polly/faqs/ · General section, 'What are Speech Marks?', checked 27 October 2026

**Captions and subtitles** · Partly — It emits sentence and word timing data as a stream your app can use for karaoke-style highlighting, but the vendor describes it for animation and highlighting, not as downloadable caption or subtitle files.

> Speech Marks are delivered in form of a JSON stream -- specifically, a set of standalone JSON objects delimited with new lines
> — https://aws.amazon.com/polly/faqs/ · General section, 'What are Speech Marks?', checked 27 October 2026

### Working with it

**Voice to video** · Partly — There is no video editor here, but it can compress a reading into a maximum duration you set so a localized track still fits the original picture.

> Amazon Polly enables you to automatically adjust the speech rate based on a maximum allotted amount of time you define with a feature called time-driven prosody.
> — https://aws.amazon.com/polly/features/ · 'Adjust the Maximum Duration of Speech' section, checked 27 October 2026

**API** · Yes — An API is the main way in, with official kits for Java, Node.js, .NET, PHP, Python, Ruby, Go, C++, iOS and Android, plus a plain HTTP interface and a command line.

> Amazon Polly can be accessed via the Polly API (and various language-specific SDKs), AWS Management Console, and the AWS command-line interface (CLI).
> — https://aws.amazon.com/polly/features/ · 'Speech Synthesis via API, Console, or Command Line' section, checked 27 October 2026

**API price** · Yes — The API price is the product price and it is public per million characters: $4.00 Standard, $16.00 Neural, $30.00 Generative, $100.00 Long-Form, with higher rates in the US government regions.

> Amazon Polly's Neural voices are priced at $16.00 per 1 million characters for speech or Speech Marks requested (when outside the free tier).
> — https://aws.amazon.com/polly/pricing/ · Pricing section, 'Pay-as-you-go model' paragraph, checked 27 October 2026

**Platforms** · Yes — web, ios, android — you can type a script and save audio in the browser console, and the vendor ships mobile kits for iPhone and Android; no desktop app for Windows or Mac is offered.

> Amazon Polly supports all the programming languages included in the AWS SDK (Java, Node.js, .NET, PHP, Python, Ruby, Go, and C++) and AWS Mobile SDK (iOS/Android).
> — https://aws.amazon.com/polly/features/ · 'Platform and Programming Language Support' section, checked 27 October 2026

**Works inside** · Partly — The named integrations are telephony and contact-center ones such as Amazon Connect, Genesys Cloud CX and the Chime kit; nothing connects it to creator tools like Canva, Adobe or WordPress.

> Amazon Polly is natively integrated with Connect Customer, AWS' cloud-based contact center solution
> — https://aws.amazon.com/polly/features/ · 'Contact center integrations' section, checked 27 October 2026

### Labelling and rights

**Who owns the audio** · Yes — The vendor states the recordings are yours and puts no limit on storing or replaying them, including static prompts replayed endlessly, with no clause ending the right if you stop paying.

> As between you and AWS, your Polly output belongs to you.
> — https://aws.amazon.com/polly/faqs/ · Data Privacy section, 'Who owns Polly recordings?', checked 27 October 2026

### Trying it, starting and getting help

**Free plan or trial** · Yes — New accounts get 5 million characters a month on Standard voices, with 1 million Neural, 500,000 Long-Form and 100,000 Generative for the first twelve months, usable on your own script in the console.

> For Amazon Polly's Standard voices, the free tier includes 5 million characters per month for speech or Speech Marks requests.
> — https://aws.amazon.com/polly/pricing/ · Free tier section, 'Millions of characters per month', checked 27 October 2026

**What the free plan leaves on your audio** · No — Free tier audio carries no stated watermark or credit requirement and the vendor advertises no restrictions on storing and reusing the output, so free characters may be published commercially.

> Amazon Polly's pay-as-you-go pricing, low cost per request, and lack of restrictions on storage and reuse of voice output make it a cost-effective way to enable speech synthesis everywhere.
> — https://aws.amazon.com/polly/faqs/ · General section, 'What is Amazon Polly?', checked 27 October 2026

**Still sold** · Yes — live — sold to new customers today, with open sign-up, a rate card updated this month and a free tier credit scheme that began in July 2025.

> With Amazon Polly, you only pay for what you use. You are charged based on the number of characters of text that you convert either to speech or to Speech Marks metadata.
> — https://aws.amazon.com/polly/pricing/ · 'Amazon Polly Pricing Details' section, checked 27 October 2026

**Sign up without a sales call** · Yes — You sign up, open the console and start typing text to hear or save audio without talking to anyone; only the custom Brand Voice route requires a sales conversation.

> Simply login to your AWS account and navigate to the Amazon Polly console (which is a part of the AWS Console). You can then use the console to type in any text and listen to generated speech or save it as an audio file.
> — https://aws.amazon.com/polly/faqs/ · General section, 'How do I get started with Amazon Polly?', checked 27 October 2026

**Contract and cancelling** · Yes — There is no term and no plan to cancel: you are invoiced each month for the characters you actually processed, so stopping use stops the charges.

> You are billed monthly for the number of characters of text that you processed.
> — https://aws.amazon.com/polly/pricing/ · Pricing section, 'Pay-as-you-go model', checked 27 October 2026

## Amazon Polly alternatives

- [Fish Audio](https://dubsheet.com/fish-audio) · Free plan, then $15 a month: Expressive text-to-speech, voice cloning and long-form studio with a credit-metered API.
- [HeyGen](https://dubsheet.com/heygen) · Free plan, then $29 a month: AI avatar video, voice cloning and video translation into 175+ languages with lip sync.
- [Narakeet](https://dubsheet.com/narakeet) · Free plan, then $6 per 30 minutes: Duration-priced browser voice-over and narrated-video maker with non-expiring capacity.
- [Murf AI](https://dubsheet.com/murf) · From $29 a month: Voice-over Studio with timeline video syncing, downloadable audio, dubbing, and a public API.
- [Voicemaker](https://dubsheet.com/voicemaker) · Free plan, then $5 a month: Credit-metered AI text-to-speech studio with 2,000+ voices, voice cloning, speech-to-speech and SRT export.
- [Typecast](https://dubsheet.com/typecast) · Free plan, then $5 a month: Expressive AI voice generator with emotion controls, voice cloning and a developer speech API.
- [Speechma](https://dubsheet.com/speechma) · Free plan, then $9 a month: Free text to speech and creator audio tools with no account or credit card.
- [TopMediai](https://dubsheet.com/topmediai) · From $35.99 a month: AI studio for text to speech, voice cloning and video dubbing on one shared credit balance.
- [Akool](https://dubsheet.com/akool) · Free plan, then $12 a seat a month: AI video suite with video translation, lip sync, text to speech, voice cloning and a voice changer.

All 9 publish a price.

[All 9 alternatives](https://dubsheet.com/amazon-polly/alternatives/)

## Sources and changes

Last changed 27 October 2026

**Sources** (4)

- https://aws.amazon.com/polly/features/, read 27 October 2026
- https://aws.amazon.com/polly/faqs/, read 27 October 2026
- https://aws.amazon.com/polly/pricing/, read 27 October 2026
- https://docs.aws.amazon.com/polly/latest/dg/limits.html, read 27 October 2026

**Updates** (1)

- 27 October 2026: Dubbing features, Free tier, License terms, Plan limits, Plans

## Is this yours?

Claim Amazon Polly: https://dubsheet.com/claim/amazon-polly

Report an error if a fact is wrong: hello@dubsheet.com, subject "Report an error: Amazon Polly (amazon-polly)".