Overview

What Is OpenASR?

OpenASR is a local-first speech-to-text project. It includes separate macOS and Windows desktop apps, a public Rust CLI, a local HTTP API, a signed model catalog, and an Apache-2.0 open core. File transcription, dictation, and live captions run on hardware you control. No account is required. There is no telemetry, no usage analytics, and no hidden cloud transcription backend. The desktop apps are closed-source products built on the open core.

Core license
Apache-2.0
Desktop platforms
macOS and Windows
Account required
No
Telemetry
None

How it works

OpenASR processes speech on hardware controlled by the user. The desktop apps, CLI, and local HTTP server all run inference locally using .oasr model packs from a signed and verified catalog.

No audio is sent to a cloud service during normal transcription. Explicit network features include model downloads, desktop update-manifest checks, optional online text refinement or translation, and optional Remote Compute. Online text processing sends recognized text—not audio—to the provider the user selects. Remote Compute sends audio only to a user-controlled device running OpenASR Server that the user has explicitly paired using TLS and trust-on-first-use certificate pinning.

Open core and desktop apps

The open core — the Rust CLI, server, and supporting libraries — is released under the Apache-2.0 license. It runs on macOS, Windows, and Linux through prebuilt releases or source builds.

The macOS and Windows desktop apps are separate closed-source products built on the open core. A native Linux desktop app is not published.

Model catalog

OpenASR provides a signed model catalog covering families such as Whisper, Qwen3-ASR, Parakeet-TDT, FireRed-AED, SenseVoice, Moonshine, Cohere Transcribe, X-ASR, and Dolphin, plus capability packs for diarization and Chinese punctuation.

Normal transcription does not silently download a missing model. The user sees and approves every model download. The CLI --offline option forbids network access and fails if a required model is missing.

Download OpenASR Get the desktop app for macOS or Windows, or install the CLI. Voice ID Name speakers on local transcripts with enrolled voiceprints. Offline speech to text Transcribe on macOS and Windows without a cloud backend. Frequently Asked Questions Common questions about privacy, platforms, models, and internet requirements.