Introduction
Corti's foundation models are specifically designed for use in the healthcare domain. Our speech recognition models are continuously refined to ensure accuracy and efficiency. To help users navigate the available options, we have introduced a tier system that categorizes functionality and performance.
Selecting the right language model tier allows you to balance capabilities and cost while ensuring the best fit for your needs.
Corti ASR Tiers
Corti's Automatic Speech Recognition (ASR) models are available in three tiers:
Tier | Description | Primary Use Case |
Base | Enables speech-to-text with a general medicine model suitable for a variety of medical specialties. | Broad support for speech-enabled clinical documentation workflows. |
Enhanced | Optimized for medical terminology recognition across different specialties, balancing speed, functionality, and quality. | Powers ambient documentation workflows. |
Premier | Provides the highest accuracy, widest language model, and most advanced functionality, including customized commands, user-defined lexicon, and specialist terminology. | Supports robust dictation or ambient documentation workflows. |
Available Languages
The table below outlines the languages currently supported by Corti ASR models, including their availability by tier:
Language | Language Code (BCP47) | Tier Availability |
English | en | Base, Enhanced, Premier |
English (US) | en-us | Base, Enhanced, Premier |
English (UK) | en-gb | Base, Enhanced, Premier |
Danish | da | Base, Enhanced, Premier (beta) |
German | de | Base, Enhanced, Premier (beta) |
French | fr | Base, Enhanced |
Swiss-German | de-CH | Base, Enhanced (beta) |
Swedish | sv | Base, Enhanced (beta) |
Spanish | es | Base, Enhanced (coming soon) |
Dutch | nl | Base |
Norwegian | no | Base |
Italian | it | Base |
Portuguese | pt | Base |
Arabic | ar | Base (coming soon) |
For any questions about language support or the best ASR tier for your use case, please reach out to our support team.
Read more about our API documentation on docs.corti.ai