Skip to main content
Corti speech to text and text generation are specifically designed for use in the healthcare domain. Speech to text (STT) language models are designed to balance recognition speed, performance, and accuracy. Text generation LLMs accept various inputs depending on the workflow (e.g., transcripts or facts) and have defined guardrails to support quality assurance of facts and documents outputs.
The language codes listed below are used in API requests to define output language for speech to text and document generation.
  • Learn more about speech to text endpoints here.
  • Learn how to query the API for document templates available by language here.

Speech to Text Performance Tiers

Corti speech to text uses a tier system to categorize functionality and performance that is available per language:
TierDescription
BaseAI-powered speech recognition, ready to integrate with healthcare IT solutionsUp to 1,000
EnhancedBase plus optimized medical vocabulary for a variety of specialties and improved support for real-time dictation1,000-99,999
PremierEnhanced plus speech to text models delivering the best performance in terms of accuracy, quality, and latency100,000+

Language Availability per Endpoint

The table below summarizes languages supported by the Corti API and how they can be used with speech to text endpoints (Transcribe, Stream, and Transcripts) and text generation endpoints (Documents):
LanguageLanguage Code
ArabicarBase
DanishdaPremier2
DutchnlEnhanced
English (US)en or en-USPremier2
English (UK)en-GBPremier2
FrenchfrPremier2
GermandePremier2
HungarianhuEnhanced
ItalianitBase
NorwegiannoEnhanced
PortugueseptBase
SpanishesBase
SwedishsvEnhanced
Swiss Germangsw-CH3Enhanced42
Swiss High Germande-CH3Premier42

Notes:
1 Use the language codes listed above for the outputLanguage parameter in POST/documents requests. Template(s) or section(s) in the defined output must be available for successful document generation.
2 Speech to text accuracy for async audio file processing via /transcripts endpoint may be degraded as compared to real-time recognition via the /transcribe and /stream endpoints. Further model updates are in progress to address the performance limitation.3 Use language code gsw-CH for dialectical Swiss German workflows (e.g., conversational AI scribing), and language code de-CH when Swiss High German is spoken (e.g., dictation).4 For Swiss German /stream configuration: Use gsw-CH for primaryLanguage as you transcribe dialectical spoken to written Swiss High German, and use de-CH for the facts outputLanguage.


Languages Available for Exploration

The table below summarizes languages that, upon request, can enabled with base tier functionality and performance.
Corti values the opportunity to expand to new markets, but we need your collaboration and partnership in speech-to-text validation and functionality refinement.Please contact us to discuss further.
LanguageLanguage Code
Bulgarianbg
Croatianhr
Czechcs
Estonianet
Finnishfi
Greekel
Hebrewhe
Japaneseja
Latvianlv
Lithuanianlt
Maltesemt
Mandarincmn
Polishpl
Romanianro
Russianru
Slovakiansk
Sloveniansl
Ukrainianuk

Language Translation

  • Translation (audio capture in one language with transcript output in a different language) is not officially supported in the Corti API at this time.
  • Some general support for translation of transcripts in English to facts in other languages (e.g. German, French, Danish, etc.) is available in stream or extract Facts requests.
  • Additional translation language-pair combinations are not quality assessed or performance benchmarked.

Please contact us if you are interested in a language that is not listed here, need help with tiers and endpoint definitions, or have questions about how to use language codes in API requests.