Skip to content
Model infoModelsSnapshot: 15 Oct 2026

Speechmatics / Speechmatics local STT

Speechmatics On-Device (2026)

Distilled and quantized native speech model for offline desktop applications, licensed to enterprise developers.

restrictedClosed weightsTranscription
Official documentation (opens in a new tab)Reviewed 13 Oct 2026
8 of 18 specification fields collected
Context windowNot collectedAwaiting source review
Maximum outputNot collectedAwaiting source review
Input billingNot collectedSpeechmatics licensed deployment
Output billingNot collectedNo inferred price

Specifications

Provider-documented facts. “Not collected” means we have not verified that field yet, not that the feature is unsupported.

Context / input limit
Not collected
Maximum output
Not collected
Reasoning
Not collected
Tool calling
Not collected
Structured output
Not collected
Fine-tuning
Not collected
Knowledge cutoff
Not collected
Release date
Not collected
Parameters
Not collected
Architecture
Distilled and quantized from the Standard cloud architecture; proprietary native inference librarySource for Architecture (opens in a new tab)
Languages
55+ languages; exact packaged language inventory depends on agreementSource for Languages (opens in a new tab)
Context caching
Not collected
Deployment options
macOS 14+ on Apple M1 or newer (CoreML); Windows 11 on Intel/AMD/ARM with DirectML and at least 2 GB GPU VRAMSource for Deployment options (opens in a new tab)
Hardware requirements
Provider estimates roughly one CPU core and 800 MB system memory plus a local AI acceleratorSource for Hardware requirements (opens in a new tab)

Access & pricing

Prices belong to an endpoint and billing tier. A provider listing does not guarantee access for every account or region.

Speechmatics licensed deployment

Provider source (opens in a new tab)
API model ID
Not collected

No pinned checkpoint selector is established.

Access
Early access by request for enterprise ISVs/OEMs; native C/C++ integration and commercial license required.
Regions
End-user device

Endpoint pricing has not been collected. Check the provider for applicable rates.

Data & enterprise terms

Transcription runs in the embedding application on the end-user device; the provider states audio and transcripts do not reach its servers. Application telemetry and downstream integrations remain controlled by the developer.

Read the service policy (opens in a new tab)

Caching

Caching can reuse earlier input processing. Availability, lifetime and billing depend on the endpoint. A cache hit still generates a new response.

01 / INITIAL INPUTProcess & retain

The system processes eligible input and retains reusable state.

02 / LATER REQUESTCheck available state

Reuse depends on matching input, session state and the cache policy.

03 / CACHE HITReuse & generate

Reusing state can reduce repeated work. Any charges follow the billing rules below.

Speechmatics licensed deployment

Not verified

No provider prompt-cache contract collected. Custom vocabulary, cached generated audio files and local Docker image caching do not establish model prompt caching.

How it works
Not collected

No provider prompt-cache contract collected. Custom vocabulary, cached generated audio files and local Docker image caching do not establish model prompt caching.

Minimum reusable input
Not collected

No provider prompt-cache contract collected. Custom vocabulary, cached generated audio files and local Docker image caching do not establish model prompt caching.

Lifetime & refresh
Not collected

No provider prompt-cache contract collected. Custom vocabulary, cached generated audio files and local Docker image caching do not establish model prompt caching.

How to enable it
Not collected

No provider prompt-cache contract collected. Custom vocabulary, cached generated audio files and local Docker image caching do not establish model prompt caching.

Usage fields
Not collected

No provider prompt-cache contract collected. Custom vocabulary, cached generated audio files and local Docker image caching do not establish model prompt caching.

Billing rules
Not collected

No provider prompt-cache contract collected. Custom vocabulary, cached generated audio files and local Docker image caching do not establish model prompt caching.

Limits & data boundaries
Not collected

No provider prompt-cache contract collected. Custom vocabulary, cached generated audio files and local Docker image caching do not establish model prompt caching.

No verified cache rate for this endpoint. An empty rate is not a zero charge.

What the measurements say

No benchmark observations have been added for this model. We do not turn a vendor description into a performance score.

Latency and throughput depend on the host, request size, reasoning settings and load. There is no measured speed or global usage figure in this record.

The record behind the model

Why it is included

Speechmatics identifies these as its own models or software, retains the software intellectual property and prohibits alteration/reverse engineering. Licensed containers and a native library do not grant an open model-weight license; SDK source availability is separate.

Classification evidence (opens in a new tab)

Sources reviewed