Abstract
This project presents a Voice-Based SQL Query Generator that enables users to interact with relational databases using spoken natural language commands instead of manually writing SQL queries. The system captures the user’s voice input and converts it into text using a speech recognition module. The recognized text is processed using a hybrid Natural Language Processing approach that combines machine learning-based intent classification with rule-based extraction of database attributes and conditions. Based on the detected intent and extracted parameters, the system automatically generates the corresponding SQL query and executes it on a SQLite database. The results are then displayed through a Flask-based web interface. The proposed system simplifies database access, reduces dependency on technical SQL knowledge, and provides a user-friendly method for retrieving and managing data through voice interaction.
Keywords
Natural Language Processing, Voice Query System, Speech Recognition, Text-to-SQL, Machine Learning, SQLite Database, Flask Web Application, Automated SQL Generation









Introduction
Relational databases are widely used in modern organizations for storing structured data such as employee records, customer information, financial transactions, and operational details. Accessing this stored information requires the use of Structured Query Language (SQL), which is the standard language used for performing operations such as retrieving, inserting, updating, and deleting data. Although SQL is powerful, it requires users to remember specific syntax rules, table structures, and column names. For non-technical users, writing SQL queries manually can be difficult and time-consuming.
With the advancement of artificial intelligence technologies, Natural Language Processing (NLP) has enabled computers to understand and interpret human language. NLP allows users to communicate with systems using simple sentences instead of structured programming syntax. Speech recognition technology further enhances usability by enabling voice-based interaction with computer systems.
This project proposes a Voice-Based SQL Query Generator that allows users to interact with relational databases using spoken natural language commands. The system converts user speech into text, analyzes the text using NLP techniques, generates the corresponding SQL query automatically, executes it on a relational database, and displays the results through a web interface. This approach simplifies database access and reduces dependency on technical SQL expertise.
Problem Statement
Traditional database systems require users to manually write SQL queries to access stored data. This process demands knowledge of SQL syntax, database schema, and logical filtering conditions. Many users, especially beginners or non-technical staff, find it difficult to remember correct SQL commands and structures. Even minor syntax mistakes can lead to execution errors or incorrect results.
Existing database interfaces provide graphical tools or text-based query systems, but these still require technical understanding of database attributes and relationships. Some NLP-based systems allow users to type queries in natural language, but most of them lack support for voice-based interaction and automatic intelligent interpretation of complex conditions.
Therefore, there is a need for a system that allows users to interact with relational databases using spoken natural language. Such a system should automatically interpret user queries, identify the intended operation, extract required conditions, and generate executable SQL statements without requiring manual coding.
Objectives
The main objective of this project is to develop a voice-based database interaction system using Natural Language Processing techniques.
The specific objectives include:
• To capture user voice input through a microphone interface
• To convert spoken audio into textual form using speech recognition
• To preprocess and normalize user queries for accurate analysis
• To implement a machine-learning model for identifying SQL query intent
• To extract database attributes, conditions, and values from natural language input
• To automatically generate syntactically correct SQL statements
• To execute generated queries on a relational database
• To display results through a user-friendly web interface
• To provide secure login authentication for database access