Speechmatics

Speechmatics is a speech API platform for teams building transcription, voice-agent and text-to-speech workflows. It is best suited to developers and organizations handling multilingual or specialized audio. Its range of models and deployment choices is a strength, but usage-based pricing and plan-specific limits call for careful selection.

Overview

Founded in 2006 and headquartered in Cambridge, United Kingdom, Speechmatics focuses on speech services that can process recorded files or live audio. Its API covers speech-to-text, voice agents and text-to-speech, making it a better fit for products and business workflows than for someone seeking a standalone personal transcription app.

Speech-to-text supports more than 55 languages, including conversations that switch languages mid-sentence. Speechmatics says its real-time service can return results in less than one second. Speaker diarization and custom vocabulary can help preserve speaker attribution and specialized terms in transcripts.

Key features

The model range addresses different workloads rather than offering one transcription rate for every need. Melia 1 handles 56 languages, mixed-language transcription, code-switching and speaker diarization at $0.12 per hour of pre-recorded audio. Oak 1 Medical costs $0.15 per hour and targets clinical language, including drug names, dosages, abbreviations and procedure terms. Enhanced costs $0.38 per hour and offers the highest accuracy, per-transcription language selection, a custom dictionary and advanced punctuation. Standard, at $0.23 per hour, has the same features as Enhanced at a lower rate but lower accuracy, making it the more defensible option for high-volume, cost-sensitive work.

For voice-agent and contact-center projects, Speechmatics names integrations with LiveKit, Vapi and Pipecat, as well as Vonage and Jambonz. Its stated use cases also include live captioning, healthcare, legal transcription and meeting platforms. These integrations and focused models make it relevant to teams embedding speech into a service; they do not make it a general-purpose meeting app.

Security provisions include ISO/IEC 27001:2022 accreditation, GDPR and HIPAA compliance, and SOC 2 Type II certification. Speechmatics says data is encrypted at rest with AES 256 and in transit using TLS 1.2 or higher, with model-training data logging off by default for self-serve and enterprise customers. Those provisions matter for sensitive workloads, although buyers still need to match deployment and support to their requirements.

Pricing

Speechmatics uses freemium, usage-based pricing and custom enterprise pricing. The Free plan costs $0.00 per free and includes $100 in credits, requires no credit card, and allows two concurrent real-time sessions and 10 pre-recorded files per second. Community, portal chat and email support are included. The small real-time concurrency cap is the main constraint for teams testing several live streams at once.

Text to Speech is $0.01 per once on a usage basis, with the first 1 million characters free and natural, low-latency synthesis aimed at voice AI. Pre-recorded transcription is charged by the hour: Melia 1 is $0.12, Oak 1 Medical is $0.15, Standard is $0.23 and Enhanced is $0.38 per hour. These model choices let buyers trade specialized language handling or stated accuracy against cost; they do not provide a flat monthly bill.

Enterprise has custom pricing and includes unlimited scale, a dedicated customer success manager and solutions engineer, plus flexible deployment options. Self-serve support is community, portal chat and email. No free trial is offered, so the free plan is the way to assess the service before paying. Self-serve accounts can have up to 50 concurrent real-time sessions, except Free accounts, which are limited to two; enterprise concurrency dynamically adjusts to demand.

Platforms

Speechmatics supports API, Linux, macOS, self-hosted, web and Windows. Deployment choices include SaaS, multi-region cloud, on-premises containers and virtual appliances, on-device, and hybrid cloud plus on-premises, with availability varying by plan. That flexibility can suit teams with infrastructure requirements beyond SaaS, but the desired deployment should be confirmed against the selected plan.

Who it's for

Speechmatics is a strong candidate for developers and organizations building multilingual transcription, live captioning, voice agents, contact-center analytics, healthcare or legal workflows. Its mixed-language support, medical model and speaker labeling capabilities address needs that a basic dictation tool may not. Individuals looking for a ready-made note-taking interface, or teams that need predictable flat-rate costs, should consider other approaches.

Pros and cons

• Pros: More than 55 supported languages and mid-sentence switching suit multilingual conversations.
• Pros: Multiple transcription models, including a clinical option, give teams a way to match terminology and accuracy needs to workload.
• Pros: SaaS, self-hosted, on-device and hybrid deployment options can accommodate different infrastructure preferences, depending on plan.
• Cons: Usage-based per-hour and per-character prices make spend depend on volume rather than a fixed subscription.
• Cons: Free accounts are limited to two concurrent real-time sessions, which can constrain live testing or small-scale deployment.
• Cons: Deployment options vary by plan, so not every customer can assume access to every hosting mode.

Alternatives

Notta is worth considering for users who want a freemium service with Android and iOS apps as well as desktop and browser options; its Free plan includes one seat and 120 transcription minutes per month.

Superwhisper may suit someone seeking voice-to-text in any app, meeting recording and transcription, and a free plan with a free trial; it supports Android, iOS, macOS and Windows.

MacWhisper is a more relevant alternative for Apple-device users who want transcription and subtitle export on a free-forever plan, with a one-time Pro plan priced at €64 a month.

Amazon Transcribe is another API option, with a free tier of 60 minutes per month for 12 months; unused monthly usage does not roll over.

AssemblyAI offers $50 in free audio credits, with five new streaming connections per minute and five concurrent pre-recorded transcriptions, an alternative for teams comparing API credit allowances and concurrency.

Google Cloud Speech-to-Text is another paid API option, including standard and dynamic batch recognition plans.

Voice In is aimed at dictation in popular sites, with built-in punctuation commands and 60 minutes of daily dictation on its free plan.

Rev AI is another API-based paid alternative.

For browsing categories, see Speech Recognition Software, Audio Transcription Software, Speech-to-Text Software or Podcast Transcription Software.

Verdict

Choose Speechmatics if you are building a speech-powered product or business workflow and need multilingual transcription, specialized models or deployment beyond standard SaaS. Its strongest case is the combination of language coverage, model choice and infrastructure flexibility. Look elsewhere if you need a standalone transcription workspace, a simple fixed subscription, or confidence that a particular deployment mode is included in a lower-tier plan.