Free ASR

Free ASR is a speech recognition client and protocol specification for macOS and iOS.

Free ASR is a zero-dependency, pure-Swift real-time streaming speech-to-text client and protocol specification reverse-engineered from the Doubao Input Method for macOS. It works without requiring an account login or device registration by connecting directly to the ASR WebSocket gateway using specific hardcoded constants and Protobuf frames. The software supports native uncompressed PCM audio input and streams both real-time partial results and three-pass rescored final transcripts. It is designed for developers who want to integrate Doubao speech recognition into their own Swift applications or test live microphone capture and audio files via the command line. The repository includes a complete English protocol specification, a recovered Protobuf schema, a self-contained Swift SDK, and a command-line testing tool.

Free ASR at a glance

Best for: Software engineer
Runs on: macOS, iOS, CLI, self-hosted, open source
Pricing: Free
Released: October 2026
Category: AI audio tool
Website: github.com/nomoat/free-asr

Key features

Real-time streaming speech-to-text
Zero-dependency pure-Swift client
No account login required
Native uncompressed PCM audio support
Live microphone capture
Command-line testing tool
Server-side three-pass rescorer

Use it for

Transcribing local audio files
Capturing live microphone input
Testing speech recognition via command line
Integrating speech-to-text into Swift apps

Alternatives to Free ASR

FluidAudio — Runs local audio AI models on Apple Neural Engine.
FunASR — It supports offline, streaming, and edge deployment with multiple models.
Esp32 Pie ASR — Faster than real time streaming speech recognition on a microcontroller alone
audio.cpp — Runs modern audio models locally with high-performance, native C++ execution.
OpenLess — Keeps data local and allows the use of your own preferred ASR and LLM providers.
AssemblyAI — Industry-leading speech models deliver high accuracy and low latency for voice applications.
WhisperLive — It provides real-time speech-to-text using multiple backends like Faster Whisper and TensorRT.
Speech-to-Speech — Fully modular, swappable components with local or hosted execution options.
Macvoice — Runs locally on Apple Silicon in under 2 GB with fast response times.
Superwhisper — Transcribes speech offline and online with custom AI modes and templates.
OpenSuperWhisper — Provides real-time transcription using customizable keyboard shortcuts and multiple engines.
Argmax OSS — Runs AI speech models locally on Apple silicon.

All Free ASR alternatives