Agenttik Coordinates Parallel AI Coding Work Across Projects

Summary

Agenttik is a developer tool created to help one person coordinate multiple AI coding agents while working on several projects in parallel. Its author describes cloning repositories and assigning different agents separate branches and folders, making it possible to review and merge their changes later. The tool is intended to reduce the mental burden of switching among agents, projects, windows, and tasks. It can work with agents from different providers and lets users choose between personal and work subscriptions for different projects. Users can review changes inside the tool, delegate more fully when desired, and retain control by setting the level of agent autonomy. Agenttik also includes shortcuts for navigation and supports scheduling or queuing tasks so agents can work sequentially. The author suggests adding project-level prompts or instructions in AGENTS.md or CLAUDE.md, including guidance for regular commits, branch use, and tests that help agents self-correct. They recommend assigning larger tasks to larger models and reviewing or iterating on each completed task. The project was built from scratch in one week, and the author says it has already helped accelerate development of the tool and other projects, while acknowledging that the models performed much of the underlying work.