Skip to main content
Corti’s text generation is purpose-built for clinical workflows. Two core capabilities — fact extraction and document synthesis — work together to turn transcripts, dictations, and existing documents into clinically reliable, source-grounded notes. Both are available via a small, consistent set of API endpoints designed for real-time and asynchronous integration patterns. The two core capabilities in the platform:

FactsR — clinical fact extraction

FactsR extracts clinically relevant facts from ambient conversations. Proven to increase downstream note completeness while keeping notes concise and source-grounded.

Guided Documents for Clinical Synthesis

Transform transcripts, extracted facts, and existing documents into precise, structured, and schema-controlled clinical documentation.

What sets Corti’s textgen apart

Compared to the many Ambient Scribe solutions that offer inflexible, limited APIs and general-purpose LLMs unfit for the healthcare domain, Corti’s platform offers clinical-grade, easy-to-build-with and flexibly designed capabilities and endpoints that enable you to build exactly the workflow that serves your customers best.

Beyond just note generation

More than simple note generation or template-based documentation — don’t just output text but structure it for enhanced control, EHR field mapping, and downstream structured-data pipelines, analytics, and CDS.

Flexibility

Generate notes from multi-source inputs — transcript, facts, existing documents – not just an ambient transcript.

Control

Guided synthesis rules to steer precise clinical documentation. Schema-driven with typed outputs (list, dict, integer, boolean, etc.) instead of free-form text strings only.

Embedded in Corti's Clinical Orchestration Platform

Ready for agent-enhanced workflows that enrich clinical documentation outputs. Ready for agent-driven workflows that create the relevant sections for EHR fields on the fly.

End-to-end pipeline — open at both ends

Upstream: pair with Corti’s standard-setting Speech-to-Text for clinical-grade transcription, or bring transcripts from any source — the platform is open. Downstream: use outputs to enrich via Medical Coding and other structured-data workflows.
Read more about reasons to choose and build on the Corti API in general.

What are you looking for?

New to Corti textgen

Start with the Beginner’s guide for a walkthrough of the concepts, then explore Core Capabilities to see what’s possible.

Specific capability guidance

Dive into the core guides — Manage Templates in Console, Sections & Templates, Guided Documents, or the Prompting Cookbook. Or browse by endpoint below.

Get started quickly

Jump straight to the Quickstart and make your first API call.

Corti Text Generation Endpoints

Streams

Real-time fact extraction to power intelligent ambient documentation and decision support workflows

FactsR™

Stateless REST endpoint to extract facts from text input

Guided Document Synthesis

Beta
Generate structured, schema-controlled documents on POST /documents — four template-supply paths, interaction-optional

Sections & Templates

Beta
Author, version and publish your own sections and templates as first-class, UUID-addressable resources

Documents Classic

Original POST /interactions/{id}/documents flow with templateKey and inline section overrides. Covers Classic templates & sections too. Planned for deprecation — see the migration guide.

Endpoint Functionality

ConnectionRESTRESTWSSREST
Data processingSynchronousSynchronousSynchronousSynchronous
ArchitectureStateless or Stateful1StatefulStatefulStateless
Text GenDocumentsDocumentsFactsRFactsR
Based on TemplatesYesYesNoNo
Quality Assurancecoming soonGuardrailscoming soonN/A
1 Guided Documents is interaction-optional: supply context (text / transcript / facts) for a stateless call, or supply interactionId to pull facts and transcripts attached to an existing interaction implicitly. The two are mutually exclusive today; combining them is on the roadmap.
Please review the languages page to learn more about languages supported per endpoint and language code to use in API requests.
Please contact us if you are interested in features that are not listed here, need help determining the best text generation endpoint for your needs, or have questions about how to configure your API requests.