This is not a graveyard of abandoned repositories. It is a place to reconstruct what actually happened: what ran, what failed, what was superseded, and what still contains useful ideas.
A working chronology built from surviving repositories, documentation, deployment traces, and project relationships.
The repository that now hosts this site was created while a broader API and self-hosted environment was taking shape.
The site documented a FastAPI backend with auth, WebSockets, MariaDB, monitoring, Docker and MicroK8s, along with live API and documentation endpoints.
A custom GPT called App Architect and Coder was configured to invoke the backend, making the model an interface into infrastructure you controlled.
Sentient Proxy explicitly referenced both jc-web-api and the broader Sentient AI system, revealing that several repositories were parts of one expanding architecture rather than isolated experiments.
The goal now is not to resume everything. It is to identify the real objective behind each artifact and decide what deserves preservation, extension, or closure.
A project only becomes useful history if its state is described accurately.
Logs, commits, deployments, API calls and observed behavior carry more weight than plans or generated documentation.
The repository name may change while the underlying problem remains the same. The objective matters more than the tool.
A completed learning experiment can stay complete. Recovering history does not obligate restarting it.