Architecture

Understanding the components that power CQL Studio

Application component Required dependency Optional integration

Drag to pan · hover or click a component or connection for details

CQL Studio UI

Browser SPA: CQL IDE, in-browser ELM translation, FHIR loaders, terminology browser, export packaging, SQL-on-FHIR (PGlite), and AI chat surface.

Core Components

CQL Studio UI

Browser IDE with in-browser ELM translation, FHIR tooling, export packaging, and SQL-on-FHIR.

CQL Studio Server

API and BFF for OIDC sessions, teams/workspaces, MCP tools, proxies, and the OpenCode gateway.

SSO IdP (OIDC)

Required OpenID Connect identity provider for sign-in — any OIDC-compatible IdP.

Application Database

PostgreSQL for users, teams, workspaces, settings, and AI session archives.

FHIR Endpoints

Four roles — evaluation, data, terminology, content — often one server via shared base URL.

OpenCode & LLM

Optional AI runner behind the server gateway; models via Ollama, OpenAI, or compatible APIs.

Cross-Platform and Hardware Support

Deploy locally or in the cloud on x64 and arm64, online or fully offline.

Deployment

Run on your machine or host on AWS, Azure, or Google Cloud with the same tooling.

CPU Architectures

Supports x64 and arm64, from laptops to data-center servers.

GPU Support

Accelerate AI workloads with Apple Silicon, NVIDIA, or AMD GPUs.

Connectivity

Works online or air-gapped behind corporate firewalls — no mandatory cloud.