Skip to main content

Compatibility

General Compatibility​

Iris (also known as Pyris) ships as part of the EduTelligence suite, and every service in the suite shares one version number. Each EduTelligence version targets a specific Artemis release line.

warning

Run matching versions from the same release cycle. Pairing an Artemis release with an EduTelligence version outside the row below is untested: features may be missing, and the Artemis-to-Pyris request contract may not match.

Version Compatibility Matrix​

Artemis VersionEduTelligence VersionStatus
8.0.x1.0.x✅ Stable
8.1.0 - 8.2.41.1.x✅ Stable
8.3.0 - 8.3.41.3.x✅ Stable
8.4.0 - 8.4.41.4.x✅ Stable
8.5.x1.5.x✅ Stable
8.6.0 - 8.6.11.6.0✅ Stable
8.6.2 - 8.6.41.6.2✅ Stable
8.7.0 - 8.7.11.7.0✅ Stable
8.7.2 - 8.7.41.7.2✅ Stable
8.8.x1.8.0✅ Stable
9.0.x2.0.x✅ Stable
9.1.x2.1.x✅ Stable
9.2.x2.2.x✅ Stable
9.3.x2.3.x✅ Stable
9.4.x2.4.x✅ Stable
9.5.x2.5.x✅ Stable
9.6.x2.6.x✅ Stable
9.7.x - 9.9.x2.7.x✅ Stable
10.0.x3.0.x✅ Stable

The same table is maintained in the EduTelligence README; update both when cutting a release.

Choosing an Image Tag​

Released versions are published to the GitHub Container Registry as ghcr.io/ls1intum/edutelligence/iris:<version> (for example ghcr.io/ls1intum/edutelligence/iris:3.0). Pin production deployments to an explicit version rather than latest, so that an upgrade is a deliberate change and a rollback is a single tag change. See Deployment.

Requirements​

RequirementDetails
Python3.13 or higher
DockerRequired for running Weaviate (vector database)
LLM API accessAt least one supported LLM provider (see below)

Supported LLM Providers​

ProviderStatusNotes
OpenAIRecommendedBest tested and most widely used in production
Azure OpenAISupportedEnterprise deployments with Azure compliance requirements
OllamaExperimentalLocal model inference — useful for development and privacy-sensitive environments
tip

For production deployments, OpenAI is the recommended provider. Azure OpenAI is a solid alternative for institutions that require Azure-based infrastructure. Ollama support is experimental and intended primarily for local development or research setups.

Deployment Options​

Iris can be deployed in several ways:

  • Docker Compose — simplest setup for development and small deployments
  • Kubernetes — recommended for production environments
  • Local development — run directly with Poetry for contributing to the project

For detailed setup instructions, see the Administrator Guide.