Sundial: Agent-Native Document Editor (www.sundial.md)

Sundial has been launched as a collaborative document editor designed to enhance the interplay between AI agents and human users in knowledge work. As AI continues to take on more tasks traditionally performed by humans, the need for transparency in what these agents produce has become critical. Sundial allows multiple agents, such as Claude Code or Codex, to work alongside humans in real-time on shared documents written in various formats like markdown and LaTeX. The platform emphasizes version control, enabling users to track every change made by both human and AI collaborators, thus ensuring a clear audit trail of the contributions. The significance of Sundial for the AI/ML community lies in its innovative approach to keeping humans at the center of the editing process while leveraging the efficiency of AI. This collaborative workspace not only addresses the growing gap between AI efficiency and human review capabilities but also provides users with an accurate understanding of what has been scrutinized. By offering a granular review panel, Sundial empowers users to maintain oversight and control over their work. Early implementations with research teams have already led to the production of conference-quality papers, showcasing Sundial's potential to redefine collaborative knowledge work in the age of AI.