DollhouseMCP
Open-source AI customization through modular elements
Blog
Engineering lessons, product context, and real stories from building DollhouseMCP in public.
Giving an AI more power means trusting it less, not more. Here is how DollhouseMCP is built so that a model being wrong, compromised, or manipulated is never enough to cause harm.
A walkthrough of the local DollhouseMCP console: guided setup, portfolio browsing, live logs, metrics, and permissions.
A reintroduction and a project announcement - how a simple prompt catalog evolved into something much bigger
The story of how we used DollhouseMCP's agent orchestration to create comprehensive documentation across three repositories in a single afternoon
How a 'critical' CI hanging issue that affected 500+ commits turned out to be a non-problem saved by a simple timeout. A cautionary tale about AI agents, assumptions, and the importance of understanding your system.
How we built an enterprise-grade MCP server while commuting, traveling, and working from coffee shops using just an iPhone
A deep dive into debugging and fixing npx/CLI execution detection for Model Context Protocol servers