Amazon Polly
- Free tier
- Trial only
- Limits tracked
- 0
- Last checked
- Last change
- None recorded
About Amazon Polly
Lexicons to fix how words are read, and speech marks for viseme and word timing.
Amazon Polly is text-to-speech on AWS, used from the console and the CLI/API. Pronunciation lexicons (W3C PLS) replace how particular words, such as company names or technical terms, are read (lexicons are stored per region). Speech marks return sentence, word, viseme and SSML-mark timing as JSON, which helps line up captions or mouth movement with the audio. Speech marks are returned instead of the audio, so they are requested separately. The free tier lasts only 12 months from your first request; there is no ongoing free tier.
Characteristics
- Pronunciation lexicons (W3C PLS) replace how specific words are read, stored per region
- Speech marks return sentence, word, viseme and SSML-mark timing as JSON
- Plain text or SSML input (pauses, rate, pitch, phonetic pronunciation, whispering)
- Usable from the AWS console and the CLI/API
Summarized by TierDiff from the vendor's own documentation, checked on Oct 16, 2026. Sources: Amazon Polly Developer Guide — Available voices, Amazon Polly Developer Guide — Generating speech from SSML documents, Amazon Polly Developer Guide — Managing lexicons, Amazon Polly Developer Guide — Speech marks, Amazon Polly — Pricing
What the free trial includes
No free-tier limits are recorded for this service.
Amazon Polly next to other text to speech
The same limits for 10 services in this category. Amazon Polly is the highlighted row.
| Service | Characters per month | Credits | Usage credit | Team members | Source |
|---|---|---|---|---|---|
| Amazon Polly AWS speech with lexicons and speech marks (viseme timing); free tier lasts 12 months. This page Trial only | not published | not published | not published | not published | pricing.us-east-1.amazonaws.comchecked Oct 23 |
| Azure AI Speech Speech synthesis with fine SSML control of pitch, rate and pronunciation; custom voices. | 500K | not published | not published | not published | azure.microsoft.comchecked Oct 23 |
| Cartesia Speech API built around response speed for voice agents; one model covers 44 languages. | not published | 20K | $1 | not published | cartesia.aichecked Oct 23 |
| CoeFont Japanese-centered voice synthesis: 10,000+ voices incl. voice actors, or record your own. | 800 | not published | not published | 1 | coefont.cloudchecked Oct 23 |
| ElevenLabs Speech generation that can shape regional accents; voice cloning and Voice Design. | not published | 10K | not published | not published | elevenlabs.iochecked Oct 22 |
| Fish Audio Voice AI platform with text to speech, voice cloning and speech to text in a web app, API and SDKs. Not read automatically | not published | not published | not published | not published | fish.audio |
| Google Cloud Text-to-Speech Speech API with voice families chosen by use; JPY billing and a Tokyo region available. | 4M | not published | not published | not published | cloud.google.comchecked Oct 23 |
| Murf AI Voiceover editor plus an API; Japanese voices with rate, pitch and pause controls. Not read automatically | not published | not published | not published | not published | murf.ai |
| Speechify Reader app that reads text, documents and web pages aloud; a separate API covers Japanese voices. Not read automatically | not published | not published | not published | not published | speechify.com |
| VOICEVOX Zundamon and other character voices, free on your own PC; commercial use with a credit. Free software | not published | not published | not published | not published | voicevox.hiroshiba.jpchecked Oct 16 |
Embed
A badge with the free-tier value, updated whenever TierDiff rebuilds. It links back to this page.
[](https://tierdiff.com/p/amazon-polly/)<a href="https://tierdiff.com/p/amazon-polly/"><img src="https://tierdiff.com/badge/amazon-polly.svg" alt="Amazon Polly free tier on TierDiff" height="20"></a>Other plans, as listed
Prices and limits of the paid plans, copied from the same page. The free tier above is what TierDiff tracks first.
Generative
| Limit | Listed value | Source |
|---|---|---|
| Price per 1M charactersPay-as-you-go, per million characters | $30 | pricing.us-east-1.amazonaws.comchecked Oct 23 Show the quote…$30 per million characters for… |
Long-Form
| Limit | Listed value | Source |
|---|---|---|
| Price per 1M charactersPay-as-you-go, per million characters | $100 | pricing.us-east-1.amazonaws.comchecked Oct 23 Show the quote…$100 per million characters for… |
Neural
| Limit | Listed value | Source |
|---|---|---|
| Price per 1M charactersPay-as-you-go, per million characters | $16 | pricing.us-east-1.amazonaws.comchecked Oct 23 Show the quote…$16 per million characters for… |
Standard
| Limit | Listed value | Source |
|---|---|---|
| Price per 1M charactersPay-as-you-go, per million characters | $4 | pricing.us-east-1.amazonaws.comchecked Oct 23 Show the quote…$4 per million characters… |