Help & Documentation
Technical documentation for the DataSpark application components and workflow.
Application Overview
DataSpark is a web application designed to demonstrate how a large language model (LLM) can act as a "natural language query layer" on top of various data sources, such as CSV files and external APIs.
The core principle is to allow users to ask questions in plain English, have the AI analyze the request, select the appropriate data source, translate the query into a machine-readable format, fetch the data, and present it back in a human-friendly way.
Key Features:
- DataBot Chat Interface: A conversational UI for interacting with data sources using natural language.
- Dynamic Data Sources: Users can upload their own CSV files (session-only) or configure persistent connections to external APIs via the "Sources" page.
- AI-Powered Query Orchestration: A series of Genkit AI flows work together to understand user intent, select data sources, translate queries, and format results.
- Manual Testing & Debugging: A "Testing" page allows developers to bypass the full conversational flow and directly test how the AI translates a query and what data is returned from a specific source.
- Comprehensive Logging: Detailed logs for each step of the AI and data fetching process are available on the "Logs" page for easy debugging.