Scholarlink

This work presents an AI-assisted environment that helps developers design REST APIs using Large Language Models. The system integrates retrieval-augmented generation with OpenAPI’s interface to deliver a structured, dialogue-based design process. Developers can define endpoints, describe functionality, and generate specifications, while the assistant searches for similar existing APIs to promote reuse. The approach is evaluated using complex, domain-specific specifications from a real-world project to assess its ability to produce valid and complete API components. The solution has been applied to the needs of a large financial institution to create compliant APIs for operations processing, enabling rapid identification of existing internal endpoints and avoiding duplicated work. The aim is to demonstrate practical value in supporting developers through faster design cycles and more consistent specifications in industrial environments.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

AI-Assisted REST API Design with Large Language Models

  • 会议论文

  • 发表日期：

    2026年

查看全文
  • Jorge Martinez-Gil,
  • Christoph Daxerer,
  • Mario Winterer,
  • Cornelia Neumüller,
  • Matthias Krump

摘要

This work presents an AI-assisted environment that helps developers design REST APIs using Large Language Models. The system integrates retrieval-augmented generation with OpenAPI’s interface to deliver a structured, dialogue-based design process. Developers can define endpoints, describe functionality, and generate specifications, while the assistant searches for similar existing APIs to promote reuse. The approach is evaluated using complex, domain-specific specifications from a real-world project to assess its ability to produce valid and complete API components. The solution has been applied to the needs of a large financial institution to create compliant APIs for operations processing, enabling rapid identification of existing internal endpoints and avoiding duplicated work. The aim is to demonstrate practical value in supporting developers through faster design cycles and more consistent specifications in industrial environments.