Skip to content

Amazon Polly vs Google Cloud Speech

Amazon Polly and Google Cloud Speech are the two big hyperscaler cloud voice suites, but Polly is a text to speech API with speech to text sold separately as Amazon Transcribe, whereas Google Cloud Speech covers both text to speech and speech to text under one brand. PyAI publishes this page and is left out of the capability table.

Prices checked October 2026·By Shaambhav Shankar

Amazon Polly

4.4 G2, 80 reviews

Text to speech API, Neural and generative voices.

$4 per 1M characters standard

Google Cloud

4.5 G2, 316 reviews

Text to speech and speech to text in one suite.

$0.016/min speech to text, $4 per 1M chars TTS

PyAI

No G2 profile yet

Voice infrastructure, one engine for the call.

$0.05/min Omni, phone line $0.01/min

Key takeaways

Amazon Polly vs Google Cloud Speech in five lines

Ratings

Google Cloud Speech scores 4.5 on G2 from 316 reviews, while Amazon Polly scores 4.4 from 80. Both are well rated, and Google's larger base reflects its speech to text product carrying the profile.

Scope

Polly does text to speech only, and AWS sells speech to text as a separate product, Amazon Transcribe. Google Cloud Speech ships text to speech and speech to text under one brand on one console.

Pricing

Polly and Google text to speech list prices track each other closely, $4 per 1 million characters standard and $16 neural on both. For speech to text Google reads $0.016 a minute, whereas Amazon Transcribe reads about $0.006 a minute in batch.

Shape

Both are raw cloud APIs. You get an audio file or a transcript back, then wire up the phone line, turn taking and barge in yourself, so a real time call is a project rather than a product.

The gap

Neither one carries the phone call on one engine. The 8 kHz phone tuned real time stack with telephony on a single bill sits elsewhere, such as PyAI at $0.05 a minute plus a $0.01 line.

Which one fits

Pick your situation, get the honest answer

Every outcome below is also written out in plain text underneath, so nothing here is hidden behind the picker. Pick one option from each row, and the answer reads the first outcome that matches all three. Several outcomes point to Amazon Polly, several to Google Cloud Speech, and the rest to one engine from PyAI.

What you need

Where you run

Monthly volume

Best fit

Pick an option on the left

The answer updates as each option is chosen.

6Capabilities both state
5Where they diverge
2Worth a trial before choosing
See which option fits each scenario
  1. PyAI. A live phone call needs one engine that listens, thinks and speaks in real time, so Omni runs at $0.05 a minute plus a $0.01 phone line with no platform fee. Hear is 8 kHz phone tuned at a 4.76% word error rate, and it is OpenAI compatible for developers.
  2. Amazon Polly. If you only need audio files and you are on AWS, Polly reads $4 per 1 million standard characters and $16 neural, with 5 million standard characters free every month.
  3. Google Cloud Text-to-Speech. On Google Cloud the list price matches Polly at $4 per 1 million standard characters and $16 for Neural2, with 4 million standard characters free a month, so pick the cloud you already run.
  4. Google Cloud Speech-to-Text. Google Speech-to-Text reads $0.016 a minute standard and $0.003 in dynamic batch, with the first 60 minutes free a month and a 316 review G2 profile behind it.
  5. Amazon Transcribe. Polly does not transcribe, so on AWS the speech to text is Transcribe, about $0.006 a minute in batch, and new accounts get 60 free minutes monthly for year one.
  6. Google Cloud, one suite. Google Cloud bundles text to speech and speech to text under one brand and one console, whereas on AWS the two sit in separate Polly and Transcribe products.
  7. Amazon, across two products. You combine Polly for synthesis and Transcribe for recognition, therefore expect two services and two line items rather than a single speech bill.
  8. Amazon, for the AWS discounts. At steady volume the AWS commitment and savings plans can trim the per character and per minute rates, so model your real monthly usage first.
  9. The pick depends on the cloud and the job. Polly and Google match on text to speech list prices, Google unifies speech to text in the same suite, and PyAI suits teams that want the whole real time call on one engine.

Pricing

What each one costs, and what moves the number

Polly and Google both price text to speech per million characters, whereas speech to text is billed per minute. PyAI charges one rate for the whole call.

Amazon Polly

Amazon Polly pricing page, October 2026.

Standard text to speech$4 per 1M chars

Neural $16, generative $30

Long-Form voices read $100 per 1 million characters. Speech to text is a separate product, Amazon Transcribe, about $0.006 a minute in batch. AWS counts 1 million characters as roughly 23 hours of audio.

5 million standard characters free every month.

Google Cloud

Google text to speech and speech to text pages, October 2026.

Speech to text standard$0.016/min

TTS $4 per 1M chars, dynamic batch $0.003/min

Text to speech matches Polly at $4 standard and $16 Neural2 per 1 million characters, Chirp 3 HD reads $30. Speech to text drops to $0.004 a minute at the top volume tier.

60 speech to text minutes free a month, 4 million TTS characters free.

PyAI

PyAI pricing page, October 2026.

Whole call per minute$0.05/min

No platform fee

Omni handles listening, the reply and the voice on one engine. Speak alone is $0.04 a minute and Hear is $0.001 a minute. Phone line $0.01 a minute, or keep your own numbers.

$50 free credit, no credit card.

Ask all three the same three questions: whether the rate is per character or per minute once synthesis and recognition are both counted, how much the free tier actually covers, and whether the phone line and turn taking are included or left to you.

Feature matrix

What each suite states it does

Each row records what the vendor publishes on its own pricing and product pages, read in October 2026. A blank means not stated, never assumed absent. PyAI is left out because it is the voice engine on the call, not a cloud catalog of speech APIs.

Amazon Polly and Google Cloud Speech compared on stated capability
CapabilityAmazon PollyGoogle Cloud
Text to speech API✓✓
Speech to text in the same suiteVia Transcribe✓
Published per character TTS rate✓✓
Neural or generative voices✓✓
Published per minute STT rateVia Transcribe✓
Always free monthly tier✓✓
SSML markup support✓✓
Custom or branded voice build✓✓
Real time streaming synthesis✓✓
Phone tuned real time call stackNot statedNot stated
Large public G2 review base80 reviews✓
Compare with PyAIGoogle pricing ↗

Deep dive

Where the two actually differ

Three areas where Polly and Google's speech suite split a shortlist.

How the suite is shaped

Amazon Polly

Text to speech only, with recognition in a separate product, Amazon Transcribe, so a full voice job spans two AWS services.

Google Cloud

Text to speech and speech to text under one brand and one console, which is why its G2 profile carries 316 reviews.

PyAI

Not in the table. One engine for listening, thinking and speaking, reached through an OpenAI compatible API or a no code builder.

What running it costs

Amazon Polly

States 7 of the 11 capabilities, $4 per 1 million standard characters, $16 neural, with Transcribe about $0.006 a minute alongside.

Google Cloud

States all 11, $4 per 1 million standard characters for TTS and $0.016 a minute for speech to text, dropping to $0.004 at volume.

PyAI

One rate of $0.05 a minute for the whole call plus a $0.01 phone line, no platform fee.

What reviewers flag

Amazon Polly

Voice realism and API integration praised, while cost, limited customization and thin docs draw complaints, across 80 G2 reviews.

Google Cloud

Transcription accuracy and real time results praised, while cost and complexity draw complaints, across 316 G2 reviews.

PyAI

No reviews yet. PyAI does not self host, and Agents is newly launched.

What reviewers say

Ratings and reviewer sentiment

Scores and tags read by hand from each vendor's G2 profile in October 2026. Google Cloud Speech holds 316 reviews against 80 for Amazon Polly, so its score rests on a wider base, while both profiles put cost at the top of the complaint list.

Amazon Polly

G2
4.4 (80)
Top praise
Voice realism
Top complaint
Expensive

Praised for: quality, voice realism, affordable pricing, API integration.

Criticized for: expensive at scale, error handling, limited customization, poor docs.

Capterra adds a 3.9 score from 10 reviews, flagging hard to brand voices and a robotic edge.

Read the profile G2

Compare with PyAI

Google Cloud

G2
4.5 (316)
Top praise
Transcription accuracy
Top complaint
Expensive

Praised for: ease of use, transcription accuracy, real time transcription.

Criticized for: expensive, pricing issues, accuracy on hard audio, complexity.

The 316 review base is the largest here, so these tags point to a settled pattern.

See all 316 reviews on G2

Google pricing ↗

PyAI

G2
No profile yet
Own benchmark
~390ms voice to voice
Build
Code or no code
No user reviews yet. PyAI does not self host and Agents is newly launched. A free sandbox key with $50 of credit and no credit card puts an AI voice on a test call before paying.
Get a sandbox key →

Decision table

If this is your situation, pick this

Read across the row.

If this is youPickWhy
Already building on AWSAmazon PollyNative TTS, $4 per 1M chars
Already building on Google CloudGoogle CloudTTS and STT on one console
You mostly need transcriptionGoogle Cloud$0.016/min, 316 G2 reviews
You mostly need voice filesEitherList prices match at $4 and $16
Lowest batch recognition rateGoogle CloudDynamic batch $0.003/min
Generative voice qualityAmazon PollyGenerative voices $30 per 1M chars
One rate for the whole callPyAIOmni $0.05/min plus $0.01 line
OpenAI compatible migrationPyAIChange the base URL, keep the SDK

Both fall short

What neither one does

Neither Amazon Polly nor Google Cloud Speech carries the phone call. Both hand back an audio file or a transcript, so neither publishes one all in rate for a live call that already includes the phone line, turn taking and barge in. PyAI runs its own engine at $0.05 a minute plus a $0.01 line, 8 kHz phone tuned.

Neither

A phone tuned real time call on one bill

Capability rows record published claims, not test results.

Verdict

So which one should you buy?

Amazon Polly is the AWS native text to speech API, while Google Cloud Speech unifies text to speech and speech to text with a larger review base. Three honest outcomes below, including the one where the answer is one engine from PyAI.

Go Amazon Polly

Amazon Polly's own page states Neural and generative voices at $4 and $16 per 1 million characters, with Transcribe alongside for recognition, for teams already on AWS.

Polly pricing ↗

Go Google Cloud Speech

Google Cloud Speech's own pages state text to speech at $4 per 1 million characters and speech to text at $0.016 a minute, under one brand with 316 reviews behind it.

Google pricing ↗

Go PyAI

The calls need one engine at one rate. PyAI runs the whole call at $0.05/min plus a $0.01 line, 8 kHz phone tuned, code or no code.

Get a sandbox key →

FAQ

Amazon Polly vs Google Cloud Speech, common questions

Is Amazon Polly or Google Cloud Speech better?

Neither overall. Amazon Polly is a text to speech API on AWS, with recognition sold separately as Transcribe, whereas Google Cloud Speech bundles text to speech and speech to text under one brand. Pick the suite that matches the cloud you already run. Teams that want the whole real time call on one engine can also look at PyAI.

Is Google Cloud Speech cheaper than Amazon Polly?

For text to speech they match closely, $4 per 1 million standard characters and $16 neural on both. For speech to text Google reads $0.016 a minute standard and $0.003 in batch, while Amazon Transcribe reads about $0.006 a minute, so price your real monthly volume on each before deciding.

Does Amazon Polly do speech to text?

No. Polly only synthesizes speech from text, priced per million characters. For transcription on AWS you use a separate product, Amazon Transcribe, which reads about $0.006 a minute in batch with 60 free minutes a month in the first year. Google Cloud Speech keeps both halves under one brand instead.

Which has more reviews, Amazon Polly or Google Cloud Speech?

Google Cloud Speech, by a wide margin. The Google G2 profile holds 316 reviews at 4.5, while Amazon Polly holds 80 at 4.4. The Google profile is the speech to text product, which draws the larger base. Both are well rated, and both list cost as the top complaint.

Do Amazon Polly or Google Cloud Speech run a phone call?

No. Both hand back an audio file or a transcript, so you still wire up the phone line, turn taking and barge in yourself. PyAI runs its own engine, Omni, at $0.05 a minute plus a $0.01 phone line, 8 kHz phone tuned, so the whole call sits on one bill.

Can I move from these APIs to PyAI without a rewrite?

Often, yes. PyAI is OpenAI compatible, so whisper-1 maps to Hear for speech to text and the /v1/audio/speech route is a drop in for synthesis. You change the base URL and keep your SDK. PyAI does not self host, and Agents is newly launched, so check those limits first.