A Dockerized deployment of Oracle Database 23ai Free with APEX 24, running on a Contabo free-tier VPS — proof that modern Oracle no longer means Oracle-license pricing. The stack now serves internal analytics dashboards with vector search and Select AI (natural-language queries).
Oracle 23ai brings AI vector search and GenAI directly into the database engine — but most teams still think "Oracle" means a six-figure procurement cycle. We wanted to prove that a small ops team can run 23ai productively on a modest cloud budget.
The internal analytics use case was straightforward: dashboards over group-wide operational data, plus a "ask a question in English" surface built on Oracle's Select AI.
Native VECTOR datatype and similarity search functions. No external Pinecone/Milvus needed — vectors live in the same DB as your operational data.
Natural-language querying: analysts type "top 10 SKUs by margin last quarter" and Oracle synthesises the SQL, runs it, and returns results.
Model data relationally, expose it as JSON — no ORM layer needed for API endpoints.
Low-code visual dashboards drawn directly on DB views. Interactive Reports, charts, and drilldowns without a JS framework.
Bounded memory, tuned SGA/PGA sizing, and workload throttling to keep the whole thing inside 2 CPU / 2 GB RAM.
docker-compose up brings the whole stack — DB, APEX, ORDS, nginx-proxy — with named volumes for data durability across container rebuilds.
We'll help you evaluate whether Oracle 23ai fits your use case, and stand up a working proof-of-concept on your infrastructure in under a week.
Book a discovery call→
