Development of a natural language to structured API data processing system with a smart home use case demonstration

This project presents the development of NL2API, a Natural Language to Structured API Data Processing System that enables intuitive human-machine communication by converting natural language (text and voice) commands into executable API calls. Leveraging advancements in Large Language Models (LLMs), the system detects user intent and translates it into structured workflows that operate across diverse APIs. A profiling framework provides an abstraction layer to accommodate heterogeneous API definitions, allowing straightforward integration into multiple domains. We evaluated several on-premises and cloud-based LLMs to identify optimal configurations for accuracy, multilingual capability, and performance. To illustrate its effectiveness, the system was deployed in a smart home environment, automating control of various IoT devices through natural language commands. By bridging the gap between human language and machine-oriented interfaces, this work demonstrates the transformative role of NLP in simplifying and broadening access to complex, API-driven systems.