International Journal of Scientific Research and Engineering Development

Efficient code review procedures are essential in modern software development in order to guarantee software quality, security, and maintainability. Traditional approaches to code review are usually time-consuming, rely on the expertise of the reviewers, and are unable to give thorough feedback with respect to a range of quality aspects. While static analysis tools can detect pre-defined coding problems, they generally do not have a good understanding of the context and do not offer specific recommendations. In order to overcome these shortcomings, this paper presents ReviewSphere AI, a multi-agent automated code review framework driven by Large Language Models (LLMs). The system makes use of dedicated AI agents which separately examine the source code for security vulnerabilities, code quality problems, and performance inefficiencies. These agents carry out their tasks at the same time and produce structured findings that are combined into a single review report including severity levels, explanations, suggestions for improvement, and overall quality scores. The framework is built using FastAPI for the backend, React for the interactive user interface, and Ollama for local LLM inference, thus allowing for a secure and cost-efficient deployment without having to depend on cloud-based AI services. The progress of the review is made available in real time via Server-Sent Events (SSE), which improves transparency and the user experience throughout the analysis process. Because of its modular design, the framework can also incorporate further review agents, which makes it scalable and capable of adapting to changing software engineering needs. By combining parallel multi-agent analysis with intelligent language models, ReviewSphere AI offers a comprehensive, context-aware, and extensible approach to automated code review, helping developers to create more secure, efficient, and maintainable software while reducing the amount of manual review required.

Suresh, Yashas R Gowda, Tejas A, Yashwanth L, Bhavya B V, "Intelligent Multi-Agent AI System with Predictive Analysis" International Journal of Scientific Research and Engineering Development, V9(4): Page(1398-1405) July-August 2026. ISSN: 2581-7175. www.ijsred.com. Published by Scientific and Academic Research Publishing.