Voyager is the AI platform for clinical data abstraction, combining frontier models, clinical guardrails, secure infrastructure, and native EHR integration to deliver fast, accurate results with minimal IT burden.
Voyager runs on Anthropic’s Claude, one of the world’s most advanced large language models. Rather than building a smaller, isolated and proprietary model in-house, we leverage billions of dollars in Anthropic's AI investment and apply that scale, accuracy, and continuous innovation to the complexity of clinical data abstraction.
Voyager layers proprietary clinical intelligence on top of frontier AI to solve problems general-purpose models cannot.
Improved temporal reasoning
It reconstructs clinical timelines across encounters and applies registry-specific lookback rules to interpret dates and events correctly.
Reduced hallucinations
Outputs are constrained by registry logic and validated against source evidence to minimize unsupported inferences.
Source-level citations
Every data element includes direct references to supporting documentation in the chart, enabling fast validation and audit defensibility.
The result is AI purpose-built for accuracy, traceability, and regulatory confidence in clinical abstraction.
Atlas empowers your workflow with faster, accurate and lower-cost clinical registry abstraction & seamless submission
AI assistant for fast answers to registry questions. Complements your existing submission workflows, lowers costs and improves efficiency.
Navigator is an Analytics solution transforming clinical data into actionable intelligence for quality improvement and accelerated research.
Short on abstraction resources? Rely on our dedicated team of 200+ seasoned clinicians for accurate and timely clinical data abstraction services. With over 200 years of collective abstraction experience and 1000+ years of clinical practice, we bring unparalleled real-world insight to your complex data.
Only our Hybrid Intelligence delivers the precise abstraction results you need. By combining cutting-edge AI with deep clinical expertise, we ensure unmatched accuracy and a nuanced understanding of healthcare data.
Voyager is built on AWS Bedrock, providing enterprise-grade infrastructure purpose-built for secure generative AI in healthcare.
All processing occurs within a governed, HIPAA-aligned cloud environment. Patient data is encrypted in transit and at rest, logically isolated by customer, and never shared with or retained by model providers. Your data does not leave your controlled boundary to train external models.
By leveraging Bedrock’s managed foundation model framework, Voyager combines frontier AI performance with the security, compliance, and operational controls health systems require.
Voyager integrates both structured and unstructured clinical data through native, standards-based FHIR connectivity, allowing it to access the full patient story without complex data mapping or traditional ETL pipelines.
Prebuilt EHR applications are designed for rapid deployment. In most environments, a single IT resource can install, configure, test, and move to production in a day. Because the integration is standards-based and does not rely on fragile custom interfaces, it remains resilient through routine EMR upgrades and version changes.
The result is faster time to value, minimal IT lift, and a durable integration model built for the realities of enterprise healthcare..
Voyager uses AI to intelligently identify, filter, and assemble the precise clinical evidence required for registry abstraction directly from the source record. Rather than relying on rigid data mappings, traditional ETL pipelines, or manually maintained codesets, Voyager interprets structured and unstructured data in context across notes, labs, reports, and encounters.
There are no ongoing interface rewrites or codeset upgrades to manage. Because Voyager reasons over the clinical record itself, it adapts as documentation patterns and workflows evolve.
Your IT team connects Voyager once through standards-based integration. From there, it continuously interprets new data in its native form, reducing technical overhead while preserving accuracy and scalability.
Carta Healthcare undergoes an independent SOC 2 audit annually to validate the strength of our security, availability, and confidentiality controls. We execute HIPAA-compliant Business Associate Agreements with every customer and technology partner, ensuring clear accountability across the entire ecosystem.
Customer data is never co-mingled. Each environment is logically isolated, with strict access controls, role-based permissions, and full audit logging. All data is encrypted in transit and at rest using healthcare-grade standards, and managed within a secure cloud infrastructure designed for regulated clinical workloads.
Security is not an afterthought. It is foundational to how Voyager is built, deployed, and operated.
Voyager is Carta’s core AI ingestion and orchestration engine that connects via FHIR, processes structured and unstructured data, and generates registry-ready answers with evidence.
Voyager connects through FHIR APIs, using read-only access without custom ETL, mapping, or system changes.
No. Voyager operates with read-only access and never modifies or writes data back into your EHR.
Each customer is deployed in a dedicated AWS account with AES-256 encryption at rest and TLS 1.2+ in transit.
No. Customer data is never used to train external models and is not exposed to third parties.
Models are hosted via AWS Bedrock, ensuring technical separation between the platform and model vendors.
No. Voyager operates within a Hybrid Intelligence model where AI suggestions are validated by experienced clinical abstractors.
Voyager processes FHIR resources including Patient, Encounter, Observation, DocumentReference, Medication, Procedure, Condition, and more.
Voyager runs in a single-tenant AWS environment with elastic scaling, encrypted storage, and monitored workloads to support high-volume abstraction without performance degradation.
Built on AWS EKS with horizontal scaling, disaster recovery across regions, and managed service support with defined SLAs.
