PaperDebugger: A Plugin-Based Multi-Agent System for In-Editor Academic Writing (arxiv.org)

PaperDebugger introduces a groundbreaking plugin-based, multi-agent system designed to enhance academic writing directly within editors like Overleaf. Unlike existing external writing assistants, this system embeds large language model (LLM) capabilities into the writing environment, facilitating deeper interaction with document states, structures, and revision histories. The technical implementation of PaperDebugger involves significant challenges, including ensuring bidirectional synchronization with the editor, managing version control, state security, and enabling multi-agent communication. These challenges are addressed through a Chrome-approved extension, a Kubernetes-native orchestration layer, and a Model Context Protocol (MCP) that incorporates features like literature search and document scoring. The launch of PaperDebugger is significant for the AI/ML community as it represents a shift towards more integrated, context-aware writing tools, allowing users to engage in localized edits and structured reviews without leaving their writing environment. It showcases advanced functionalities, such as parallel agent execution and diff-based updates, while maintaining a user-friendly interface. Early analytics indicate strong user engagement, underscoring the potential for this in-editor assistant to transform the academic writing process, making it more efficient and less disruptive. This development paves the way for future enhancements in academic workflows leveraging AI technology.