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.
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 Version | EduTelligence Version | Status |
|---|---|---|
| 8.0.x | 1.0.x | ✅ Stable |
| 8.1.0 - 8.2.4 | 1.1.x | ✅ Stable |
| 8.3.0 - 8.3.4 | 1.3.x | ✅ Stable |
| 8.4.0 - 8.4.4 | 1.4.x | ✅ Stable |
| 8.5.x | 1.5.x | ✅ Stable |
| 8.6.0 - 8.6.1 | 1.6.0 | ✅ Stable |
| 8.6.2 - 8.6.4 | 1.6.2 | ✅ Stable |
| 8.7.0 - 8.7.1 | 1.7.0 | ✅ Stable |
| 8.7.2 - 8.7.4 | 1.7.2 | ✅ Stable |
| 8.8.x | 1.8.0 | ✅ Stable |
| 9.0.x | 2.0.x | ✅ Stable |
| 9.1.x | 2.1.x | ✅ Stable |
| 9.2.x | 2.2.x | ✅ Stable |
| 9.3.x | 2.3.x | ✅ Stable |
| 9.4.x | 2.4.x | ✅ Stable |
| 9.5.x | 2.5.x | ✅ Stable |
| 9.6.x | 2.6.x | ✅ Stable |
| 9.7.x - 9.9.x | 2.7.x | ✅ Stable |
| 10.0.x | 3.0.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
| Requirement | Details |
|---|---|
| Python | 3.13 or higher |
| Docker | Required for running Weaviate (vector database) |
| LLM API access | At least one supported LLM provider (see below) |
Supported LLM Providers
| Provider | Status | Notes |
|---|---|---|
| OpenAI | Recommended | Best tested and most widely used in production |
| Azure OpenAI | Supported | Enterprise deployments with Azure compliance requirements |
| Ollama | Experimental | Local model inference — useful for development and privacy-sensitive environments |
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.