---
title: "On-device and on-premises speech recognition for regulated teams"
description: "Transcribe and dictate without sending audio to a cloud API. Phonon runs on CPUs, Apple silicon and NVIDIA GPUs inside your environment; Detta dictates on the Mac's Neural Engine. Open weights, measured accuracy."
canonical: "https://www.fermionresearch.com/deploy/on-device-speech-recognition/"
source: "Fermion Research"
---

Deploy/facts checked 13 November 2026

# On-device and on-premises speech recognition for regulated teams

If audio cannot leave the building, the model has to come to the audio. A 156 MB model that runs on a laptop or a CPU server makes that practical.

[Talk to us](https://www.fermionresearch.com/deploy/)contact@fermionresearch.com

## Two ways to keep audio inside

- On the device: Detta, our dictation app for the Mac, recognises speech on the Mac's Neural Engine. With Wi-Fi off, dictation keeps working. Detta asks for a Google sign-in, and sharing dictations to improve our models is on by default with an off switch in Settings; with sharing off, no dictation is uploaded.
- On your servers: fermion serve runs Phonon-2.5 (25 languages, 156 MB) or Phonon-2.5 English on CPU or GPU machines in your cloud or data centre, with an OpenAI-compatible endpoint and an API key.

## Accuracy you can check

The models are open and scored on public test sets: Phonon-2.5 English averages 4.82 % word error on the seven Open ASR Leaderboard sets; Phonon-2.5 averages 11.71 % across 25 languages on FLEURS. Your team can re-run the numbers on your own recordings before deploying.

## Capacity on ordinary hardware

Published Phonon-2 figures (Phonon-2.5 shares its size and architecture): 143 times realtime on eight AMD Zen 5 cores, 174 times realtime on an M5 MacBook Air GPU, 6,680 times realtime batched on an H100. A single eight-core CPU server transcribes more than a hundred hours of audio per hour.

## Custom vocabulary

Clinical, legal and financial speech is full of terms a general model has not seen. The fermion runtime takes a hotword list per call (names, drugs, products) that biases recognition toward the listed terms without changing the transcript when they are not said; Detta has a dictionary for names and terms.

```text
fermion transcribe phonon-2 visit.wav --hotwords @drug-names.txt
```

## Questions

## Related

- [Self-host a speech-to-text server](https://www.fermionresearch.com/models/phonon-2-5/self-host/)
- [Detta, free dictation for the Mac](https://www.fermionresearch.com/products/detta/)
- [Replace API model calls with a small model](https://www.fermionresearch.com/deploy/replace-api-models-with-small-models/)

## Sources and method

- Phonon numbers: Phonon-2.5 release post (2026-10-22); throughput: Phonon-2 model page. Hotwords: /docs/speech/ (fermion-research 0.2.15).
- Detta behaviour: Detta privacy policy and the 2026-10-04 Wi-Fi-off test (lane detta_vs_wispr_20261004).
- No BAA or contract claim is made; FOUNDER_ITEMS.md.

Canonical: https://www.fermionresearch.com/deploy/on-device-speech-recognition/

Source: [https://www.fermionresearch.com/deploy/on-device-speech-recognition/](https://www.fermionresearch.com/deploy/on-device-speech-recognition/)
