We may earn a commission if you buy through links on this page. It never changes a score or a ranking. How we make money
The bottom line
Strong multilingual speech APIs, with usage-based costs and an API-focused setup.
Speechmatics is worth a look if your small team is building transcription or voice features, especially for multilingual or privacy-sensitive work. Its free plan includes $100 in credits, and Pro lists starting rates of $0.12 per hour for pre-recorded audio and $0.15 per hour for real-time audio.
This is an API rather than a ready-made team transcription workspace. Seats are unlimited on Pro, but your bill depends on audio use, and subscription prices aren't published. Teams that want a simple, fixed per-seat bill should compare other options.
Good fit if
- Small teams building voice products
- Multilingual transcription workflows
- Privacy-sensitive audio applications
Skip it if
- Teams wanting fixed seat pricing
- Nontechnical teams needing a finished app
- Occasional users after free credits run out

ScoresHow it rates
What works
- Supports more than 55 languages
- Real-time transcription under one second
- Includes speaker diarization and custom vocabulary
- Flexible cloud, on-premises, and on-device deployment
- Pro seats are unlimited
What doesn't
- Pro subscription prices aren't published
- Costs vary with audio usage
- Free access is limited to $100 in credits
- API setup may need developer time
- Some add-on prices are available only on request
01Who Speechmatics is for
Speechmatics fits a small product team—say, one to ten people—that needs speech recognition inside its own app or workflow. It turns audio into text through APIs, including real-time transcription and pre-recorded file processing. The product supports more than 55 languages, speaker identification, and custom vocabulary, with deployment options for cloud, on-premises, or on-device use.
The best fit is a team building voice agents, dictation, transcription, or a specialized workflow where language coverage and deployment choices matter. The homepage also describes a medical model for healthcare dictation and ambient scribing, plus HIPAA, GDPR, SOC 2 Type II, and ISO 27001 compliance or certification.
This is less suitable if you need a finished meeting-notes app or predictable per-person pricing. Speechmatics sells API usage, and its Pro subscription prices aren't published. A small team should budget by expected audio volume, not headcount.
02Speechmatics free plan and trial
Self-serve Free includes $100 in credits, two concurrent real-time sessions, and capacity for 10 pre-recorded files per second. The pricing page lists community support, portal chat, and email. It’s a useful way to try the API or handle low-volume use before paying for more audio.
The free plan is credit-based, not an unlimited free tier. Once the credits are used, the next self-serve option is Pro, with rates starting at $0.12 per hour for pre-recorded audio and $0.15 per hour for real-time audio. Speechmatics doesn't publish a free-trial period or a fixed monthly Pro subscription price. Check the credit balance and expected audio volume before moving production traffic onto a paid plan.
03What a small team pays for Speechmatics
Speechmatics charges by audio use, not by seat. Pro lists starting rates of $0.12 per hour for pre-recorded audio and $0.15 per hour for real-time audio, with unlimited seats. For a simple comparison, 10 total audio hours in a month would start at $1.20 pre-recorded or $1.50 real-time. That same usage would cost the same for one person, three people, or five people; divided evenly, it works out to $1.20 or $1.50 for one, $0.40 or $0.50 per person for three, and $0.24 or $0.30 per person for five.
At those starting rates, 10 hours each month for a year comes to $14.40 pre-recorded or $18 real-time, before any subscription terms or changes in rate. These are usage examples, not quoted subscription prices. The pricing page says Pro can be paid as you go, with credit packs, or by monthly or annual subscription, but doesn't publish subscription prices.
Add-ons can change the bill: translation is $0.49 per hour where available, while eligible pre-recorded models have chapters at $0.30 per hour, topics at $0.15, and summaries and sentiment at $0.09 each. Ask about pricing for PII redaction, entity detection, and audio alignment.
04Getting started with Speechmatics
Because Speechmatics is API-first, expect to connect it to your product or workflow rather than simply invite a team into a ready-made transcription app. The product offers real-time and pre-recorded speech-to-text, while deployment can be in the cloud, on-premises, or on-device. The right setup depends on which of those you need and how your app handles audio.
A small team with a developer can start with the Self-serve Free credits and test a low-volume workflow. If you are building a voice agent, Speechmatics founder guidance on Reddit points to Pipecat as one platform choice, but that discussion isn't a substitute for a setup plan tailored to your app.
One court reporter on Reddit said making the online portal work in real time took technical workarounds and some tech savvy. That example is not a universal setup estimate, but it’s a useful warning: teams without a developer should confirm the exact workflow and support needs before committing.
05When small teams outgrow Speechmatics
Speechmatics is more likely to hit a budget or deployment boundary than a seat limit. Pro includes unlimited seats and 50 concurrent real-time sessions, but audio usage is billed by the hour. The pricing page lists the starting rates; subscription prices and volume tiers aren't fixed publicly, so ask sales for a quote if usage is growing or predictable monthly budgeting matters.
Enterprise is the listed step up for organizations that need real-time concurrency adjusted to demand, unlimited pre-recorded files per second, a dedicated customer success manager and solutions engineer, or on-premises, on-device, and hybrid deployment options. Its price is custom.
Reddit feedback is positive in a court-reporting discussion: one working reporter recommends Speechmatics, while also saying a comparable real-time setup took technical workarounds. Another commenter said it can be worth paying per minute only when a hearing warrants it. That points to the key trade-off for a growing team: keep the API for valuable or frequent audio, but compare alternatives if occasional use and hands-on integration make the spend hard to justify.
From the threadsWhat users say
The available Reddit comments lean positive, especially from court-reporting users who like the recognition quality. They also flag practical caveats: some workflows take technical workarounds, and paying by audio use may not suit every job.
Speechmatics has been great
I’m a working reporter and I’d go speechmatics 100%
for use in court, you gotta be a bit tech savvy to get it to be comparable to eclipse
not every hearing is going to be worth paying minutes to use Speechmatics for
Questions people ask
Does Speechmatics have a free plan?
Yes. Self-serve Free includes $100 in credits, two concurrent real-time sessions, and 10 pre-recorded files per second. The plan also lists community support, portal chat, and email. It is credit-based, so continued use after those credits are spent requires a paid option.
How much does Speechmatics cost for a small team?
Pro rates start at $0.12 per hour for pre-recorded audio and $0.15 per hour for real-time audio. Pro includes unlimited seats, so the cost depends on audio volume rather than headcount. The vendor doesn't publish its monthly or annual subscription prices.
What languages does Speechmatics support?
Speechmatics says one model supports more than 55 languages, including natural language switching within conversations. Check with the vendor about language coverage and suitability for your specific use case before routing important audio through it.
Can Speechmatics run on-premises or on-device?
Yes. The homepage describes cloud, on-premises, and on-device deployments. Enterprise also lists on-premises, on-device, and hybrid options. Ask the vendor which deployment is available for your use case and what it would cost.
Is Speechmatics suitable for healthcare audio?
The homepage describes a medical model for healthcare dictation and ambient scribing, and lists HIPAA among its compliance credentials. It also says the model can cut errors on key terms by up to 50%. Confirm requirements, deployment, and pricing with Speechmatics for your workflow.
Sources
- Speechmatics homepage
- Speechmatics pricing page
- Reddit: 36 Years in Voice AI | Built One of the First Speech Systems in 1989 | Dr Tony Robinson (Founder, Speechmatics) - AMA for next 24 hrs
- Reddit: Dragon vs Speechmatics for new voice reporters?
- Reddit: Dragon vs Speechmatics for new voice reporters?
- Reddit: Case CAT and Speechmatics...thoughts?
