23AI Query Assistant FAQs

Question 1

Which core technologies does this tool utilize?

Accepted Answer

It's built with Oracle Autonomous Database (23ai), Oracle Vector Search, a locally hosted LLaMA 3.2 LLM (via Ollama), and a FastAPI server, all managed with Python and the `oracledb` driver.

Question 2

What data types can be queried?

Accepted Answer

The assistant can query both structured data (e.g., customer details) using traditional SQL and unstructured text (e.g., complaints, documents) using semantic vector search, all within Oracle Autonomous Database.

Question 3

How does it intelligently process queries?

Accepted Answer

It uses a locally hosted LLaMA 3.2 LLM (via Ollama) to interpret user questions and intelligently decide whether to generate a traditional SQL query for structured data or a semantic vector query for unstructured data.

Question 4

What is the 23AI Query Assistant?

Accepted Answer

The 23AI Query Assistant is a proof-of-concept system that converts natural language questions into SQL or semantic search queries over data stored in Oracle Autonomous Database (ADB), leveraging Oracle 23ai's vector search and a local LLaMA 3.2 LLM.

Question 5

Does it require an external vector database?

Accepted Answer

No, a key feature is that all vector data storage, embedding generation (using Oracle 23ai's built-in models), and querying are performed directly within Oracle Autonomous Database, eliminating the need for external vector databases.

23AI Query Assistant

23AI Query Assistant

Key Features

Use Cases

Key Features

Use Cases