A research team at Duy Tan University announced the launch of AgenTex v1.0 – a Multi-Agent AI System to Support Scientific Research – on the morning of May 17, 2026.

This software utilizes a Multi-Agent AI architecture – a Scientific Research Assistant – developed with the aim of establishing a comprehensive research support platform, contributing to improving the quality of international publications and promoting innovation in the academic environment.

Contributing to Solving Bottlenecks

According to SCImago 2024, Vietnam currently ranks 49th globally in terms of annual scientific publication volume, but still lags significantly behind many countries in the region such as Thailand, Malaysia, and Singapore.

Surveys at research and training institutions in Central Vietnam have identified five common bottlenecks hindering the research and publication process:

- Incorrect writing according to IMRaD academic standards;

- Lack of peer review before submission;

- Complex LaTeX formatting (67% of researchers spend more than 20% of their time on formatting - according to Overleaf 2023);

- Literature reviews taking approximately 2-5 days;

- Fragmented research tools.

Based on this reality, Duy Tan University's AgenTex was developed as a technical solution with the aim of becoming a Virtual Research Assistant to support scientists, optimizing technical processes in research, thereby helping researchers dedicate more time to academic thinking to make valuable scientific contributions.

The system is built on the core philosophy of "Human-in-the-Loop," in which:

- AI plays a supporting role, suggesting, critiquing, and optimizing technical tasks in research;

- Scientists remain at the center and have the decision-making power throughout the entire research development process through interaction with AI.

"AgenTex doesn't write for researchers, but rather acts as an intelligent companion: helping to form sharper ideas, review literature faster, write more accurately, critique more thoroughly, and present information according to international publication standards, so that researchers can focus on the irreplaceable part: scientific thinking and knowledge contribution." - Phan Vu Minh Tue, student and author of the AgenTex project, Duy Tan University, shared.

Multi-Agent AI Ecosystem

AgenTex v1.0 is built on a Multi-Agent AI architecture consisting of 5 system layers and an ecosystem of 9 specialized AI agents, operating synchronously on a single web interface.

 

Each agent takes on a specific role in the research cycle, such as: idea generation, document synthesis, draft building, peer review, consistency checking, LaTeX format standardization, etc.

AgenTex v1.0 has 3 core features:

- Drafting: Simultaneous searching of ArXiv, Semantic Scholar, OpenAlex, Tavily → IMRaD outline → abstract draft with citations → LaTeX code. Estimated 10-15 minutes for 15-20 documents.

- Peer Review: Simulates the peer review process – checking arguments, data, compliance with journal formats, and ensuring research ethics.

- Formatting: Direct LaTeX compilation in the browser, support for IEEE, Elsevier, Springer, and ACM templates, and PDF/TEX/ZIP export.

The system allows simultaneous data retrieval from ArXiv, Semantic Scholar, OpenAlex, and Tavily, supports the creation of research outlines using the IMRaD structure, simulates peer review, and standardizes articles according to IEEE, Elsevier, Springer, or ACM standards.

The research, with initial test results, involved over 50 faculty members and graduate students, conducted through research activities and the development of academic projects.

"Having directly experienced the software with graduate students and undergraduates, I believe AgenTex is not simply an editing tool, but an intelligent ecosystem supporting academic research and editing.

The system effectively supports the processing of formatting, structure, and organization of research content, thereby allowing researchers to dedicate more time to focusing on academic thinking and developing core scientific values," said Dr. Hoang Thi Huong, Acting Rector of the School of Language & Social Sciences (LHSS), Duy Tan University.

Towards a large-scale AI research platform

Following the release of the AgenTex v1.0 trial, the research team plans to expand AgenTex in the 2026-2027 period in the following directions: Integrating Multimodal, a multi-author collaboration space, and a cloud service model for universities.

With the vision of becoming an "AI university" in the near future, Duy Tan University is gradually developing an ecosystem of technology integrating Artificial Intelligence (AI) into all aspects of the university's operations, from administration and training to learning, research, and entrepreneurship.

Mr. Nguyen Son Hai - Deputy Director of the Department of Science, Technology & Information (Ministry of Education & Training), who attended the launch event, shared: "The launch of AgenTex v1.0 shows that Duy Tan University has chosen the right practical path, directly linked to science, technology, and research activities in the higher education environment - a field that is increasingly developing."