About

About DollhouseMCP

DollhouseMCP is an open-source platform for AI customization built around reusable elements, local ownership, and safety-aware execution.

Mission

The goal is simple: give people real control over AI behavior without forcing everything back into disposable prompts. DollhouseMCP turns behavior, capabilities, structure, context, and workflows into readable building blocks you can version, activate, combine, and keep.

What 2.0 focuses on

• Reusable elements

  Six core element types provide a composable model for behavior, capabilities, templates, memory, and execution.

• Local-first ownership

  Your portfolio lives with you, which makes the system portable, inspectable, and easier to trust.

• Structured MCP operations

  MCP-AQL replaces tool sprawl with semantic endpoints and introspection-based discovery.

• Runtime safety

  Gatekeeper policies, approval flows, and execution-state tracking keep agentic workflows controllable.

Who is behind it

DollhouseMCP is led by Mick Darling, founder of DollhouseMCP Inc. His background includes long-running work in natural language processing, product strategy, and applied AI systems, including earlier NLP patent work through Tomorrowish.

The public product direction spans the core DollhouseMCP platform, the Dollhouse Collection for community element sharing, and MCP-AQL as the protocol layer that makes the system easier for LLMs to use.

What makes the project different