AI-driven academic diagram generation tool that automates the creation of CVPR/NeurIPS standard illustrations from paper abstracts. It targets researchers and academics seeking to streamline the process of generating high-fidelity scientific figures, offering a structured workflow and support for multiple AI models to enhance productivity and visual quality.

How It Works

The project employs a strict "Logic (Architect) -> Vision (Renderer)" workflow. An AI logic model (e.g., GPT-5.1, DeepSeek, Gemini) analyzes input text or documents (PDF/images) to generate a structured "visual schema" or blueprint. This schema is then reviewed and potentially refined by the user, with options to incorporate reference images for style guidance. Finally, a vision AI model (e.g., Gemini-3-pro-image-preview) renders the academic diagram based on the schema, producing high-quality illustrations.

Quick Start & Requirements

• Primary install/run command: Clone the repository, navigate to the frontend directory, run npm install, and then npm run dev.
• Prerequisites: Node.js and npm are required for local development. API keys for the chosen AI models (GPT, Gemini, DeepSeek) are necessary for functionality.
• Links:
  • Live Demo: https://aia-academic-illustrator.vercel.app
  • Video Tutorial: https://www.bilibili.com/video/BV1bMm3BYEKD/
  • Issues: https://github.com/qwwzdyj/AIA-Academic-Illustrator-/issues

Highlighted Details

• Supports multiple AI models for both logic analysis (GPT, DeepSeek, Gemini) and image rendering (Gemini).
• Features browser-side PDF to image conversion using pdf.js.
• Allows users to upload reference images for style guidance during diagram generation.
• Offers a "Bring Your Own Key" (BYOK) mode for enhanced data security.
• Provides a bilingual (Chinese/English) user interface and saves history locally (up to 2 images).

Maintenance & Community

The project acknowledges original author @BAIKEMARK and community support from Datawhale. Issues and feedback can be submitted via the GitHub issues page.

Licensing & Compatibility

The project is licensed under the MIT License, permitting commercial use and modification.

Limitations & Caveats

The tool relies on external AI model APIs, requiring users to manage their own API keys and associated costs. Local storage for history is limited to two images. The effectiveness of generated diagrams is dependent on the quality of the input and the chosen AI models.