From Specification to Service: Accelerating API-First Development Using Multi-Agent Systems
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866908605893050368 |
|---|---|
| author | Chauhan, Saurabh Rasheed, Zeeshan Sami, Malik Abdul Kemell, Kai-Kristian Waseem, Muhammad Zhang, Zheying Rasku, Jussi Saari, Mika Abrahamsson, Pekka |
| author_facet | Chauhan, Saurabh Rasheed, Zeeshan Sami, Malik Abdul Kemell, Kai-Kristian Waseem, Muhammad Zhang, Zheying Rasku, Jussi Saari, Mika Abrahamsson, Pekka |
| contents | This paper presents a system that uses Large Language Models (LLMs)-based agents to automate the API-first development of RESTful microservices. This system helps to create an OpenAPI specification, generate server code from it, and refine the code through a feedback loop that analyzes execution logs and error messages. The integration of log analysis enables the LLM to detect and address issues efficiently, reducing the number of iterations required to produce functional and robust services. This study's main goal is to advance API-first development automation for RESTful web services and test the capability of LLM-based multi-agent systems in supporting the API-first development approach. To test the proposed system's potential, we utilized the PRAB benchmark. The results indicate that if we keep the OpenAPI specification small and focused, LLMs are capable of generating complete functional code with business logic that aligns to the specification. The code for the system is publicly available at https://github.com/sirbh/code-gen |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_19274 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | From Specification to Service: Accelerating API-First Development Using Multi-Agent Systems Chauhan, Saurabh Rasheed, Zeeshan Sami, Malik Abdul Kemell, Kai-Kristian Waseem, Muhammad Zhang, Zheying Rasku, Jussi Saari, Mika Abrahamsson, Pekka Software Engineering This paper presents a system that uses Large Language Models (LLMs)-based agents to automate the API-first development of RESTful microservices. This system helps to create an OpenAPI specification, generate server code from it, and refine the code through a feedback loop that analyzes execution logs and error messages. The integration of log analysis enables the LLM to detect and address issues efficiently, reducing the number of iterations required to produce functional and robust services. This study's main goal is to advance API-first development automation for RESTful web services and test the capability of LLM-based multi-agent systems in supporting the API-first development approach. To test the proposed system's potential, we utilized the PRAB benchmark. The results indicate that if we keep the OpenAPI specification small and focused, LLMs are capable of generating complete functional code with business logic that aligns to the specification. The code for the system is publicly available at https://github.com/sirbh/code-gen |
| title | From Specification to Service: Accelerating API-First Development Using Multi-Agent Systems |
| topic | Software Engineering |
| url | https://arxiv.org/abs/2510.19274 |