Speechmatics
  • Enterprise
  • Pricing
  1. Features And Deployments

A speech-to-text API with three ways to deploy it: cloud, on-prem, or on-device.

Batch and real-time transcription, speaker diarization, translation, and summarization, in one API.

Built for voice agents, contact center tools, medical transcription and more.

Processing modes

Get the very best performance whether you're transcribing pre-recorded files or live audio.

Configuration

File transcription

Process thousands of hours of pre-recorded files, whenever you need them, and fast.

Configuration

Live transcription

Transcribe media as it happens. Get initial transcriptions in milliseconds, with context-driven accuracy improvements over time.

Deployment options

Run Speechmatics wherever your infrastructure and compliance needs require

Choose how Speechmatics runs — in the cloud, on your own infrastructure, or directly on-device — deployed however suits you.

Configuration

On-Prem

Meet architecture, security and compliance needs by hosting our API in your own environment. Deploy with Docker containers, Kubernetes, or a preconfigured virtual appliance.

Configuration

Cloud

Get secure and scalable access to our API through our cloud deployment and get instant access to all our new features, languages and updates.

Configuration

On-Device

Run Speechmatics directly on your devices for ultra-low latency and maximum data privacy — nothing sent over a network. Get cloud-parity accuracy (within 5%), full speaker diarization, and sub-second latency, on Mac or Windows. Ideal for use cases where connectivity is limited and data must stay local.

Models

Every model runs on the same API. Pick the one that matches your priorities.

Melia 1: automatic language detection, labelling, and seamless code-switching in one API, across all 56+ supported languages. Available now for batch transcription, with real-time support on the way.

Model overview

Model

Best for

Real-time

Batch

Multilingual

Enhanced

Highest accuracy, one language per file or stream

✓ Yes

✓ Yes

✗ No

Standard

Strong accuracy with faster turnaround and lower cost

✓ Yes

✓ Yes

✗ No

Melia 1

Audio that switches between languages mid-conversation, across all 56+ supported languages

Coming soon

✓ Yes

✓ Yes

Feature support by model

Feature

Enhanced

Standard

Melia 1

Speaker & channel diarization

✓ Yes

✓ Yes

✓ Yes

Word-level timestamps & punctuation

✓ Yes

✓ Yes

✓ Yes

Custom dictionary

✓ Yes

✓ Yes

Not yet

Confidence scores

✓ Yes

✓ Yes

Not yet

Translation, summaries, chapters

✓ Yes

✓ Yes

Not yet

Code-switching across 56+ languages

✗ No

✗ No

✓ Yes

Diarization ships on every model, including Melia 1. It's one of the areas Speechmatics is strongest, and it's not something every multilingual model gets right.

See pricing for model costs, or the developer docs for configuration syntax.

Transcription Features

Everything you need to hit the highest accuracy possible

Our customization options allow you to finely tune your set up to achieve high accuracy with even the most unique words and phrases.

Feature

Custom Dictionary

Boost accuracy for proper nouns, acronyms or industry-specific terms by providing a list of custom words.

Feature

Speaker & Channel Diarization

Track who said what and when with speaker labelling for each word, available for both batch and real-time transcription.

Feature

Numeral Formatting

Identify and correctly format numbers, dates and currencies automatically to improve transcript readability and enable effective post-processing.

Feature

Profanity & Disfluency Detection

Aid comprehensibility and compliance by detecting and optionally removing words that are considered profanities or hesitations.

Features

File Formats

Minimize the resource needed to prepare audio or video files with support for all major audio and video formats along with automatic sample rate detection.

Advanced Features

Easily push a variety of media formats to the API

Easily push a variety of media formats to the API and get a rich set of metadata to support your post processing needs.

Features

Confidence Scores

Collect confidence scores for every word in the transcript to enable efficient human review and editing.

Features

Word Timings

Get accurate timestamps for every word in the transcript to allow for post-processing and improved end user experience.

Feature

Advanced Punctuation & Casing

Improve readability with language-specific capitalization and punctuation including commas, question marks and exclamation marks.

Features

Audio Events

Improve accessibility & fully-automate tedious captioning by identifying and labelling non-speech sounds in media, using AI.

Languages

Partner with Speechmatics to maximize your total addressable market

We deliver for multilingual, multicultural and multinational businesses, with coverage of 56+ languages across a range of dialects and accents.

Accuracy tuning extends to specific industries too, including finance terminology and a dedicated medical model reaching 93% accuracy with 50% fewer keyword errors than the next-best system.

Language Coverage

56+ languages, with code-switching at the word level.

Accents and dialects

Whether you need Brazilian Portuguese or Canadian French, we have you covered with a single language model that supports all associated accents and dialects.

Translation

Transcribe and translate audio to and from English for over 30 languages using a single API call.

Language Identification

Simplify integration and ensure accurate transcription with automatic detection of the language spoken.

Speech understanding

Transcription pairs with summaries, sentiment, topics, and chapters in the same API call.

Translation

With automatic translation with a single API call, you can translate media and provide captions for over half the world’s population.

Summaries

Instantly generate summaries for social and video platforms, so viewers know what to expect, without you having to manually write.

Sentiment

Don’t just rely on reviews. See how customers are feeling about every aspect of your service by identifying sentiment throughout calls.

Topics

Your audience don’t want to (always) watch long media. Give them the topics discussed and the timestamps so they can engage with what they are most interested in.

Chapters

As well as being divided up and summarized, each chapter is given a heading, making it super easy to find the most engaging content.

Certifications & compliance

Built to meet enterprise security and compliance requirements

Speechmatics meets the security and compliance bar enterprises need, wherever you deploy. See speechmatics.com/security for full detail.

Certifications

ISO/IEC 27001:2022

Accredited to the ISO/IEC 27001:2022 information security standard.

Certifications

SOC 2 Type II

SOC 2 Type II-certified, independently audited on security controls over time, not just a point in time.

Certifications

HIPAA

Fully compliant with the Health Insurance Portability and Accountability Act, for healthcare and clinical use cases.

Certifications

GDPR

Compliant with GDPR and other regional privacy directives.

Certifications

Encryption

Data encrypted at rest (AES-256) and in transit (TLS 1.2+).

Certifications

Data residency

Processing environments in the US, EU, or Australia, to meet data sovereignty requirements.

Resources for features and deployments

Carousel slide image
Company

Speechmatics versus Whisper: how Adobe Premiere's on-device speech engine got rebuilt

Quantization was the key to fitting a cloud-grade model on a laptop. Getting the full optimization chain to cooperate around it was the hard part.

Andrew Innes
Andrew InnesChief Architect
[alt: Orange gradient background with "Melia" centrally placed, highlighting multilingual support with code-like symbols scattered.]
Product

Introducing Melia, our new multilingual speech-to-text model

A multilingual speech-to-text model from Speechmatics, with code-switching across 56+ languages. Available today in production preview, starting with batch transcription.

Yahia Abaza
Yahia AbazaSenior Product Manger
Carousel slide image
Technical

The Adobe story: How we made cloud-grade AI work on your laptop

Behind the build: what it takes to make cloud-grade speech recognition work inside Adobe Premiere, and why Whisper raised the stakes.

Andrew Innes
Andrew InnesChief Architect
Carousel slide image
Technical

How to build a microbatching workflow with the Speechmatics API

Build a cleaner path between batch and real time. Learn when micro-batching makes sense, how to chunk audio, submit jobs, stitch JSON, and scale safely with the Speechmatics API.

Speechmatics
SpeechmaticsEditorial Team
Carousel slide image
Company

Adobe and Speechmatics deliver cloud-grade speech recognition on-device for Premiere

Adobe Premiere users can run the most accurate on-device transcription locally; efficient enough for a laptop, powerful enough for professional work.

Speechmatics
SpeechmaticsEditorial Team
Carousel slide image
Use Cases

Best speech-to-text AI guide: APIs, platforms and services compared

Speech-to-text has moved from novelty to enterprise infrastructure. Here's how the leading platforms stack up in 2026 — and how to pick the right one.

Tom Young
Tom YoungDigital Specialist
[alt: Two healthcare professionals, wearing blue scrubs, engage in conversation in a hospital in Sweden]
Product

Speechmatics launches new Swedish medical model, cutting transcription errors by 40%

Expanding a Nordic medical lineup with 3.91% KWER model that delivers sub-second latency across Swedish, Finnish, Danish, and Norwegian clinical workflows.

Yahia Abaza
Yahia AbazaSenior Product Manger
speaker diarization
Product

What is Speaker Diarization and why does it matter in voice AI?

The breakthrough technology helping AI understand conversations like humans do.

Stuart Wood
Stuart WoodProduct Manager

Ready to Understand Every Voice?

Sign up to our free speech-to-text SaaS Portal and we’ll guide you through the integration of our API.

SpeechmaticsSpeechmatics
ISO/IEC 27001
Queen's Award for Enterprise 2019
G2 High Performer badge
GDPR Compliant
HIPAA Compliant
ISO 27001 Certified
SOC 2 Certified
Product
  • Speech to Text
  • Real-Time
  • Languages
  • On-Device
  • Features and deployments
  • Enterprise
  • Voice Agent API
Use Cases
  • Medical & Healthcare
  • AI Voice Agents
  • Legal Transcription
  • Contact Center Solutions
  • Media & Captioning
  • Speech Analytics
  • Note-Taking & Meeting Assistants
  • EdTech
Pricing
  • Pricing Options
Resources
  • Documentation
  • Blog and Latest News
  • Case Studies
  • Service Status
  • Security
  • Hey AI, learn about us
  • GitHub Academy
About
  • About Us
  • Careers
  • Brand
  • Contact Us
  • Startup Program
  • Community
Compare
  • vs AssemblyAI
  • vs AWS
  • vs Deepgram
  • vs Gladia
  • vs Google
  • vs Microsoft Azure
  • vs Soniox
  • vs ElevenLabs
  • vs OpenAI
  • Speechmatics Discord
  • Speechmatics X
  • Speechmatics LinkedIn
  • Speechmatics YouTube channel
  • Speechmatics Reddit
  • Speechmatics Github
  • Privacy Policy
  • Terms of Website
  • Terms of Service
  • Cookie Policy
  • Sitemap
Copyright © Speechmatics 2026