Architecture
Understanding the components that power CQL Studio
Core Components
Web Interface
Modern, browser-based tools and integrated development environment (IDE) for writing, editing, testing, and publishing CQL code. Provides syntax highlighting, autocomplete, and real-time validation.
Embedded CQL Translator
Translates CQL code into executable formats. Handles parsing, validation, and conversion of CQL expressions into ELM.
Data & CQL Server
FHIR resource server providing persistent storage and retrieval of clinical data and CQL execution capabilities. Manages data retrieval, CQL compilation, and query processing.
Terminology Server
FHIR resource server providing access to healthcare terminologies, value sets, and code systems. Supports standard terminologies like SNOMED CT, LOINC, and ICD-10. Content is customizable and can be replaced with a custom terminology server.
AI/LLM Integration
Optional AI assistance via the Ollama model runner and integrated Model Context Protocol (MCP) services. Language model for code generation and assistance can be replaced with a custom model.
Engine Test Runner
Comprehensive testing framework for testing CQL engine implementations against official HL7 test cases. Provides detailed test report data and visualization tools for cross-comparison to other engines. Test runner can be pointed at any network-accessible CQL engine implementation.
Cross-Platform and Hardware Support
CQL Studio runs in a wide range of environments. You can deploy in local or hosted cloud environments, on x64 and arm64 CPU architectures, and leverage GPUs including Apple Silicon, NVIDIA, and AMD. It works in both Internet-accessible online environments and completely offline environments behind corporate firewalls.
Run CQL Studio on your own machine for development and testing, or deploy to hosted cloud environments such as AWS, Azure, or Google Cloud for team and production use. The same tooling runs identically in either setting.
CQL Studio supports both x64 (Intel/AMD) and arm64 (e.g. Apple Silicon M-series, ARM servers) architectures. Run on the hardware you already use, from laptops to data-center servers, without architecture lock-in.
AI and translation workloads can leverage GPUs for better performance. Supported options include Apple Silicon (Metal), NVIDIA (CUDA), and AMD GPUs, so you can use the accelerators available in your environment.
Use CQL Studio in Internet-accessible online setups or in fully air-gapped, offline environments behind corporate firewalls. No mandatory cloud dependency—deploy where your security and compliance requirements demand.
Configuration & Integration
CQL Studio is designed with flexibility in mind. Each component can be configured or replaced to meet your organization's specific needs, security requirements, and infrastructure constraints.
Connect to your own FHIR servers, databases, or data warehouses
Choose from any MCP-compatible language model or deploy your own
Run locally, on-premises, or in cloud environments
Plugin system for custom functionality and integrations