Google Cloud Speech-to-Text or AssemblyAI: two bills
Google Cloud Speech-to-Text bills $0.016 a minute for the v2 Standard model and gives the first 60 minutes free. AssemblyAI bills $0.15 and $0.21 an hour and gives 185 free hours. Transcribsi is $10 a month with no minute limit. Read on 24 October 2026.
Google Cloud bills $0.016 a minute for the v2 Standard model and gives 60 free minutes of the v1 API. AssemblyAI bills $0.15 and $0.21 an hour, with 185 free hours. Transcribsi is $10 a month with no minute limit.
Google Cloud, AssemblyAI and Transcribsi side by side
the speech-to-text API of Google Cloud, sold by the minute of audio the service processes
a voice AI platform: pre-recorded, realtime and sync speech-to-text APIs, speech understanding, guardrails and an LLM gateway
a transcription service for files: a REST API, an MCP server and plugins for Claude and ChatGPT
per minute: $0.016 for the v2 Standard model up to 500,000 minutes a month, then $0.01, $0.008 and $0.004; dynamic batch is $0.003 a minute; $0.016 a minute is $0.96 an hour
per hour: $0.21 for Universal-3.5 Pro and $0.15 for Universal-2
per month: $10 a month paid yearly ($120 a year) with no minute limit, or $20 a month paid monthly
the v1 API is free for the first 60 minutes of audio a month; the v2 Standard model has no free tier named on the page
185 hours of pre-recorded transcription and 333 hours of streaming, with no credit card
$0: 3 files a day, up to 30 minutes per file, no card
no number given on the pricing page; the supported-languages page lists the BCP-47 codes of the models
99 for Universal-2 and 18 for Universal-3.5 Pro, which also switches code natively
99+ languages, set by you or detected
the v2 API with dynamic batch, which processes audio at a lower urgency and at a discounted rate
a realtime API: 5 new streaming connections a minute on the free plan and 100 a minute on pay-as-you-go, scaling by itself
not offered: we take a file that is already finished
each channel is billed separately, so a stereo hour counts as two hours of audio
multichannel audio is billed per channel and transcribed independently, which the page says helps with multi-speaker recordings
speaker labels on request, with timecodes on every segment
medical dictation and medical conversation models at $0.078 a minute
no medical model on the pricing page; the voice agent costs $4.50 an hour and includes recordings and transcripts
no medical model, no topic model and no sentiment model
TXT, Markdown, SRT, VTT, CSV, and DOCX on Pro
a Google Cloud bill with a pricing calculator, a console and a contact-sales path
a monthly invoice for the previous month's usage, no minimum commitment, and the option to buy through the AWS Marketplace
one subscription, no meter and no commitment
medical models, seven named recognition models, a discount ladder, a console and a pricing calculator
a voice agent, guardrails, an LLM gateway, a dictation API, speech understanding and 99 languages
none of that, and no live audio, no batch mode, no console, no add-ons and 99+ languages
A discount ladder and an hourly meter
One of these is a product inside a cloud console, the other is an API company with a playground and a no-code demo. Both sell the same minute of audio, but they count it differently and they pay you back differently: Google gives you a minute of free audio once a month, AssemblyAI gives you hundreds of free hours. Our service is a flat $10 a month for finished files, with no console in front of it.
How to read two very different bills
- 1
Convert the unit before comparing
Google's v2 Standard model is $0.016 a minute, which is $0.96 an hour. AssemblyAI's Universal-2 is $0.15 an hour. The two land in the same neighbourhood once the unit is the same.
- 2
Look at the free part closely
Google's v1 API is free for the first 60 minutes of audio a month, and the v2 Standard model has no free tier named. AssemblyAI gives 185 hours of pre-recorded and 333 of streaming. Ours is 3 files a day of 30 minutes.
- 3
Watch the channel multiplier
Google bills every channel of a recording separately. AssemblyAI bills multichannel audio per channel as well, and transcribes each channel on its own. We take one file and ask for a speaker split inside the transcript.
- 4
Decide if you want a console
Google gives you a console, a pricing calculator and a contact-sales path; AssemblyAI gives you a playground and an AWS Marketplace listing. We give you an API, an MCP server and a file, and no console to learn.
Tips
A discount ladder only helps the big buyer.
Sixty free minutes are a trial, not a tier.
Channels are billed twice, once per ear.
A console is not a feature you pay for.
Questions this page answers
Which of the two is cheaper for a small volume?
Google's v2 Standard model starts at $0.016 a minute and AssemblyAI's cheapest model is $0.15 an hour. Our plan is $10 a month with no minute limit, which is the one that does not change when your volume changes.
What does AssemblyAI give away for free?
185 hours of pre-recorded transcription and 333 hours of streaming, with no credit card required. Ours gives 3 files a day of up to 30 minutes, no card, and it renews every day rather than ending.
How many languages does each one name?
AssemblyAI names 99 for Universal-2 and 18 for Universal-3.5 Pro. Google names no number on the pricing page and lists BCP-47 codes instead. We publish 99+ languages for recognition: 99+ languages.
Can Google transcribe a medical conversation?
Yes, the pricing page lists medical dictation and medical conversation models at $0.078 a minute. AssemblyAI has no medical model there, and we have none either — that is a gap on our side.
Do these services export a transcript?
Neither pricing page names export formats. We publish six: TXT, Markdown, SRT, VTT and CSV on every plan, and DOCX on Pro.
Sources
- CGoogle Cloud — Speech-to-Text pricingcloud.google.com · checked 24 October 2026
- CGoogle Cloud — supported languagescloud.google.com · checked 24 October 2026
- AAssemblyAI — pricingassemblyai.com · checked 24 October 2026
- TTranscribsi — pricingtranscrib.si · checked 24 October 2026
- WAssemblyAI — supported languageswww.assemblyai.com · checked 24 October 2026
Reading: 9 minutesUpdated Transcribsi editorial team
Drop a recording — the text arrives in a minute
Speakers by name, a time on every line, export to TXT, SRT, VTT, CSV and Markdown.