Ligth Circuit
AI-assisted project orchestration and dev history
LC is a project workspace designed for structured collaboration between a developer and AI agents. It stores projects, tasks, decisions, implementation history, and feedback in a single traceable workflow. Through MCP integration, agents can retrieve project context, work on tasks, and return structured results without losing the reasoning and decisions accumulated during development. The system focuses on human-controlled orchestration: AI assists with implementation, while task state, review, and final decisions remain under the developer’s control.
- Role
- Creator, Product Architect, System Analyst
- System
- Java 21 / Spring Boot / Thymeleaf / PostgreSQL / MCP
Development without losing context
Software development with AI often breaks down not because an agent cannot write code, but because project context is scattered across chats, task trackers, repositories, and personal notes. LC keeps projects, tasks, decisions, implementation history, and feedback in a single structured workspace. Each task preserves not only its current state, but also the path that led to it: the initial request, refinements, implementation attempts, review comments, and final result. Through MCP, AI agents can retrieve the relevant project context and return structured updates directly to LC. The developer remains responsible for review and final decisions, while the system preserves continuity between separate work sessions and different agents.