denario

Denario - AI-driven Research Workflow Automation System

Capabilities Overview

Denario is a research assistant system based on a multi-agent architecture that can automatically complete end-to-end research tasks from data analysis to paper publication.

Use Cases

1. Research Data Analysis and Hypothesis Generation

When you have a set of research data and need to discover valuable research directions, Denario can analyze dataset features and automatically generate research hypotheses and scientific questions to be tested. It supports various data types such as time series data and experimental data.

2. Research Methodology Development and Experiment Execution

When you need to design research protocols and run computational experiments, Denario can help you build structured research methodologies, run numerical computations, generate visualizations, and produce complete analysis results. It supports common scientific computing tools such as pandas, sklearn, and scipy.

3. Automated Academic Paper Writing

After completing research, when you need to write a journal paper, Denario can automatically generate a LaTeX paper that conforms to the target journal's formatting requirements, supporting various journal styles such as APS, including figure/table integration and complete LaTeX source code.

Core Features

1. End-to-end Research Pipeline

Denario provides full automation of the research workflow: data description → hypothesis generation → method design → experiment execution → paper writing. Each stage can be executed automatically or manually intervened, supporting a flexible hybrid work mode. All outputs are structured for easy version control and reproducibility.

2. Multi-agent Task Coordination

Built on the AG2 and LangGraph frameworks, Denario coordinates multiple specialized AI agents to handle different tasks: hypothesis generation experts, methodology design experts, data analysis experts, paper writing experts, etc. The agents collaborate to ensure coherence and quality of the research process.

3. Intelligent Paper Generation and Formatting

Automatically converts research outputs into LaTeX papers formatted for journal submission. Supports multiple journal style templates, automatically handling figure/table insertion, reference formatting, section structure, etc. The generated LaTeX source can be directly compiled to PDF and meets academic publishing standards.

Frequently Asked Questions

Which LLM APIs does Denario support? How to configure?

Denario supports Google Vertex AI, OpenAI, and other LLM services compatible with AG2/LangGraph. You need to configure API keys via environment variables or a .env file. For detailed configuration guidance (including Vertex AI setup) see the references/llm_configuration.md document.

Can Denario fully automatically generate papers?

Denario can automatically generate paper drafts, including research background, method descriptions, results presentation, and conclusion analysis. However, the generated papers still require human review and revision to ensure scientific accuracy, logical coherence, and quality of expression. Denario is more like an intelligent research assistant rather than a complete replacement for human researchers.

What Python environment does Denario require?

Denario requires Python 3.12 or higher. It is recommended to install using uv: uv add "denario[app]". If you encounter package dependency conflicts, use a virtual environment or Docker container. The Docker image has LaTeX and all dependencies preinstalled and is ready to use out of the box.