Welcome to GENIE.AI
The one-pager — what GENIE.AI is, who it is for, and where to go next.
What GENIE.AI is
GENIE.AI is an open-source retrieval-augmented-generation framework built for the public sector. It bundles a multilingual chat UI, a self-hosted RAG pipeline, an admin dashboard, an identity layer, and a configurable observability stack — so a government or NGO can deploy a sovereign AI assistant without sending a single query to a third-party model API.
The pipeline is built on top of OPEA, the Open Platform for Enterprise AI, and integrates with the Digital Public Goods ecosystem.
Who it is for
- Citizens — ask questions about public services in their own language.
- Operators — keep the system healthy, update the knowledge base, and scale it.
- Developers — extend the API, customize the UI, swap models.
- Contributors — improve the codebase, the docs, and the translations.
How a query flows
User → Frontend (Vue) → Backend (BFF) → ChatQnA → Embedding
→ Retriever (ArangoDB) → Reranker → LLM (vLLM) → Response
Every stage emits OpenTelemetry traces so a single query can be debugged end-to-end in Grafana / VictoriaTraces.
What is sovereign here
- Models run on your hardware (vLLM / TEI) — no model API keys required.
- Data stays in your ArangoDB and document store.
- Identity is delegated to your Keycloak (or any OIDC provider).
- Telemetry is optional — opt in with
ENABLE_OBSERVABILITY=1and everything lands in Victoria* stores you control.
Where to next?
- New user → Quickstart: User
- Installing → Quickstart: Deployer
- Extending → Quickstart: Developer
- Contributing → Quickstart: Contributor