Platform Overview

OrcAI is a self-hostable platform for governed, retrieval-backed AI assistants. It originated in the SokratesT project and still supports learning workflows where instructors curate material and students interact with a grounded AI tutor. OrcAI has since moved beyond a fixed instructor–student model: individuals and groups can build specialised knowledge bases and assistants for their own domains.

What OrcAI Does

OrcAI combines content ingestion, semantic search, and chat:

1. Workspace members upload source material.
2. Background workers process files and extract usable text or image descriptions.
3. Processed content is stored in S3-compatible object storage.
4. Extracted content is indexed in Qdrant for retrieval.
5. Users chat with the platform.
6. Relevant context is retrieved and sent to the configured AI model endpoint.

Core Areas

• Chat: user-facing AI conversations, optionally started from configured bots.
• Content library: uploaded files and processed assets used as knowledge sources.
• Bots and blocks: reusable AI behavior and knowledge configurations.
• Access control: organizations, users, groups, invitations, and resource sharing.
• Administration: providers, models, quotas, jobs, and platform configuration.

Suggested Reading

Getting Access

Access to the managed service of OrcAI is currently only available via direct invites from users already on the platform. If you want to create your own organisation (isolated workspace) to invite users yourself, contact us at sokratest@hochschule-rhein-waal.de