gpt_academic provides a modular, interactive interface for LLMs that is tailored for reading, polishing, and writing academic papers. It supports hot-reloadable function plugins, custom shortcuts, and specialized document translation tools.
Project overview
The project specializes in academic document workflows, offering specific tools for translating and proofreading Arxiv, LaTeX, and PDF papers while allowing users to execute arbitrary tasks through hot-reloadable plugins.
Project type
Infrastructure
Deployment
Refer to project documentation
License
GPL-3.0
Best for
Researchers and developers seeking an interactive web GUI for reading, polishing, and writing academic papers using LLMs.
Users needing customizable quick shortcuts and hot-reloadable plugins to execute arbitrary tasks within their LLM interface.
Key capabilities
Supports customizable function plugins that allow hot updates to execute arbitrary tasks, with low difficulty for writing and debugging.
Supports letting GPT generate flowcharts, state transition diagrams, Gantt charts, pie charts, and GitGraphs.
Allows adding new quick shortcut buttons via the UI appearance menu or through code configuration.
Limitations and risks
Official Gradio has many compatibility issues; users must install the exact Gradio version specified in requirement.txt.
Some browser translation plugins interfere with the operation of the software's frontend.
Running local large models like ChatGLM4 requires high hardware specifications, specifically at least 24G of VRAM.
Getting started
Clone the repository, configure your API_KEY in config.py, install the requirements, and run main.py. Direct Python deployment requires configuring environment variables and editing configuration files, though one-click scripts and Docker images are also provided.