Google Health AI team has released MedASR, an open weights medical speech to text model that targets clinical dictation and physician patient conversations and is designed to plug directly into modern AI workflows.<\/p>\n
MedASR is a speech to text model based on the Conformer architecture and is pre trained for medical dictation and transcription. It is positioned as a starting point for developers who want to build healthcare based voice applications such as radiology dictation tools or visit note capture systems.<\/p>\n
The model has 105 million parameters and accepts mono channel audio at 16000 hertz with 16 bit integer waveforms. It produces text only output, so it drops directly into downstream natural language processing or generative models such as MedGemma.<\/p>\n
MedASR sits inside the Health AI Developer Foundations portfolio, alongside MedGemma, MedSigLIP and other domain specific medical models that share common terms of use and a consistent governance story.<\/p>\n
MedASR is trained on a diverse corpus of de identified medical speech. The dataset includes about 5000 hours of physician dictations and clinical conversations across radiology, internal medicine and family medicine.<\/p>\n
The training pairs audio segments with transcripts and metadata. Subsets of the conversational data are annotated with medical named entities including symptoms, medications and conditions. This gives the model strong coverage of clinical vocabulary and phrasing patterns that appear in routine documentation.<\/p>\n
The model is English only, and most training audio comes from speakers for whom English is a first language and who were raised in the United States. The documentation notes that performance may be lower for other speaker profiles or noisy microphones and recommends fine tuning for such settings.<\/p>\n
MedASR follows the Conformer encoder design. Conformer combines convolution blocks with self attention layers so it can capture local acoustic patterns and longer range temporal dependencies in the same stack.<\/p>\n
The model is exposed as an automated speech detector with a CTC style interface. In the reference implementation, developers use MedASR training uses JAX and ML Pathways on TPUv4p, TPUv5p and TPUv5e hardware. These systems provide the scale needed for large speech models and align with Google\u2019s broader foundation model training stack.<\/p>\n Key results, with greedy decoding and with a six gram language model, are:<\/strong><\/p>\n A minimal pipeline example is:<\/strong><\/p>\nAutoProcessor<\/code> to create input features from waveform audio and AutoModelForCTC<\/code> to produce token sequences. Decoding uses greedy decoding by default. The model can also be paired with an external six gram language model with beam search of size 8 to improve word error rate.<\/p>\nPerformance on medical speech tasks<\/strong><\/h3>\n
\n
Developer workflow and deployment options<\/strong><\/h3>\n