Speechmatics or Deepgram: the same audio, two currencies
Speechmatics counts hours from $0.12 and gives $100 of credit. Deepgram counts minutes from $0.0048 and gives $200. Transcribsi is $10 a month with no minute limit of finished audio. Read on 24 October 2026.
Speechmatics bills by the hour: $0.12 for Melia 1, $0.23 for Standard, $0.38 for Enhanced, and $100 of free credit. Deepgram bills by the minute from $0.0048 and gives $200 of credit. Transcribsi is $10 a month with no minute limit of finished audio.
Speechmatics, Deepgram and Transcribsi side by side
an enterprise speech-to-text engine with four named models: Melia 1, Enhanced, Oak 1 for medical work and Linden 1 for voice agents
a speech-to-text API for developers, with text-to-speech, a voice agent API and audio intelligence on the same account
a transcription service for files: a REST API, an MCP server and plugins for Claude and ChatGPT
per hour: $0.12 for Melia 1, $0.23 for Standard, $0.38 for Enhanced and $0.15 for the medical model Oak 1
per minute: $0.0048–$0.0092 a minute depending on the model; $0.0048 a minute is about $0.29 an hour
per month: $10 a month paid yearly ($120 a year) with no minute limit, or $20 a month paid monthly
$100 in credits, no card required, then pay for what you use
$200 of credit on an account with no card, then pay-as-you-go
$0: 3 files a day, up to 30 minutes per file, no card
56 for Melia 1, which also switches language mid-sentence; the feature table says 55+ languages
45+ for the Nova models, with automatic language detection
99+ languages, set by you or detected
pre-recorded and real-time, with 50 concurrent real-time sessions on self-serve, 2 of them on the free tier
REST and WebSocket streaming: up to 50 REST sessions and 150 WebSocket sessions, 225 on the Growth plan
not offered: we take a file that is already finished
no size and no duration limit named on the pricing page
no size and no duration limit named on the pricing page
up to 30 minutes on the free plan, no minute limit on Pro
TXT, Markdown, SRT, VTT, CSV, and DOCX on Pro
translation at $0.49 an hour, chapters at $0.30, topics at $0.15, summaries and sentiment at $0.09, plus text-to-speech at $0.011 per 1,000 characters with the first million free
redaction at $0.0020 a minute, speaker diarization at $0.0020, entity detection at $0.0017 and keyterm prompting at $0.0013
none of that: no translation, no summaries, no topics, no sentiment, no redaction and no text-to-speech
the pricing page lists an on-device section and asks about on-premises and on-device deployment; the enterprise tier is sold on request
a dedicated EU endpoint at api.eu.deepgram.com, GDPR ready, SOC 2 Type 1 and Type 2, and a BAA for enterprise health data
our own engine outside the United States; a clean English file can still go to a US cloud provider when the fallback is on
a medical model, a voice agent model, on-device deployment, translation, summaries and sentiment
text-to-speech, a voice agent API, audio intelligence, redaction, entity detection and keyterm prompting
none of that, and no live audio, no on-premises deployment, no add-ons and 99+ languages
An hour, a minute and a month
Two engines, two currencies. Speechmatics sells to enterprises and counts the hour; Deepgram sells an API to developers and counts the minute. The free parts are credits on both sides, $100 and $200, and neither asks for a card. Our service sits outside that exchange: one flat month for finished files, with Russian among the 99+ languages we publish.
How to compare two speech engines
- 1
Convert the units first
Melia 1 is $0.12 an hour and Deepgram's cheapest Nova model is $0.0048 a minute, about $0.29 an hour. The gap narrows to a few cents, which is why the headline numbers look further apart than the bills do.
- 2
Read the free parts as credits
Speechmatics hands out $100 of credits and Deepgram $200, with no card either way. Ours is not credit: 3 files a day of up to 30 minutes, free every day, and the counter resets overnight.
- 3
Count the languages honestly
Melia 1 handles 56 languages and switches language mid-sentence, while Deepgram names 45+ for the Nova models. We publish 99+ languages, and Russian is the one our engine was tuned for.
- 4
Decide which side you are on
Both rivals are built for live audio, voice agents and enterprise deployment. We are the finished file: up to 30 minutes, six export formats and speaker labels when you ask for them.
Tips
An hour and a minute are the same audio.
Credits are a trial with a price tag.
Mixed-language audio is a separate feature.
An EU endpoint is not the same as on-premises.
Questions this page answers
Which of the two costs less per hour of audio?
It depends on the model: Speechmatics starts at $0.12 an hour for Melia 1, Deepgram at $0.0048 a minute for Nova-3 monolingual. Our plan is $10 a month with no minute limit, whichever unit you prefer to think in.
What do you get before you pay anyone?
Speechmatics gives $100 of credits with no card, Deepgram gives $200 of credit with no card. Ours gives 3 files a day of 30 minutes, and you can keep doing that for as long as you like.
Can Speechmatics handle 99+ languages at once?
Yes, Melia 1 transcribes 56 languages without choosing one up front and switches language mid-sentence. Deepgram's Nova models name 45+ languages. We handle 99+ languages only.
Which one can run inside your own network?
Speechmatics lists on-premises and on-device deployment and sells an enterprise tier on request, while Deepgram offers an EU endpoint at api.eu.deepgram.com. We run our own engine outside the United States and deploy nothing on your hardware.
Do they export a transcript to a file?
Neither pricing page names export formats. We publish six: TXT, Markdown, SRT, VTT and CSV on every plan, and DOCX on Pro.
Sources
- SSpeechmatics — pricingspeechmatics.com · checked 24 October 2026
- DDeepgram — pricingdeepgram.com · checked 24 October 2026
- TTranscribsi — pricingtranscrib.si · 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.