AI Tool

mcp-server Review

mcp-server is a Model Context Protocol (MCP) server that enables dynamic AI persona management from markdown files, allowing compatible AI assistants to activate and switch between different behavioral personas.

Visit mcp-server→

mcp-server - AI tool for server. Professional illustration showing core functionality and features.

1DollhouseMCP's mcp-server is an open-source platform released under the AGPL-3.0-or-later license.

2DollhouseMCP 2.0 introduced a complete element model, MCP-AQL, and 38 starter elements including personas and skills.

3The server facilitates dynamic AI persona management from markdown files, supporting AI assistants like Claude.

4It features five semantic endpoints for structured operations and a safety-aware execution layer for agent workflows.

𝕏 in ↑↗

⚡

mcp-server at a Glance

Best For

Pricing

freemium

Key Features

Integrations

See website

Alternatives

See comparison section

🏢

About mcp-server

Leadership

Mick Darling

Connect

𝕏

X / Twitter@mickdarling

</>Embed "Featured on Stork" Badge▼

HTML

<a href="https://www.stork.ai/en/mcp-server" target="_blank" rel="noopener noreferrer"><img src="https://www.stork.ai/api/badge/mcp-server?style=dark" alt="mcp-server - Featured on Stork.ai" height="36" /></a>

Markdown

[![mcp-server - Featured on Stork.ai](https://www.stork.ai/api/badge/mcp-server?style=dark)](https://www.stork.ai/en/mcp-server)

overview

What is mcp-server?

mcp-server is a Model Context Protocol (MCP) server developed by DollhouseMCP that enables developers, writers, professionals, and others to manage AI personas dynamically from markdown files. It allows AI assistants, such as Claude, to activate and switch between different behavioral personas. DollhouseMCP's mcp-server functions as an open-source Model Context Protocol (MCP) server, designed to enhance and customize AI (Artificial Intelligence) experiences. It achieves this by enabling dynamic management of AI personas, skills, templates, agents, and memories. The server allows AI assistants, including Claude, to interact seamlessly with its functionalities through the standardized MCP (Model Context Protocol) protocol, facilitating natural language interaction without requiring manual API (Application Programming Interface) calls. The core package, @dollhousemcp/mcp-server, is distributed under the AGPL-3.0-or-later license. This architecture supports the creation, modification, and sharing of modular building blocks for AI customization, acting as a bridge for AI models to access and utilize external tools and data through a structured protocol.

quick facts

Quick Facts

Attribute	Value
Developer	DollhouseMCP
Business Model	Open Source / Freemium Hybrid
Pricing	Open-source core, freemium model
Platforms	API, Web
API Available	Yes
Integrations	Claude (other) and other compatible AI (Artificial Intelligence) assistants
License	AGPL-3.0-or-later
Last Updated	March 16, 2026

features

Key Features of mcp-server

DollhouseMCP's mcp-server provides a comprehensive set of features for advanced AI (Artificial Intelligence) customization and management. Its architecture supports modularity and dynamic control over AI (Artificial Intelligence) behaviors, leveraging the Model Context Protocol (MCP).

1Dynamic AI (Artificial Intelligence) persona management from markdown files, enabling flexible behavioral definitions.
2AI (Artificial Intelligence) persona activation and switching for compatible AI (Artificial Intelligence) assistants, including Claude (other).
3API (Application Programming Interface) availability for programmatic interaction and integration into existing workflows.
4Creation, management, and sharing of modular AI (Artificial Intelligence) personas, skills, templates, agents, and memories.
5Customization of AI (Artificial Intelligence) behavior and capabilities across different AI (Artificial Intelligence) models and services.
6Orchestration of multi-agent workflows and management of prompt interactions for complex AI (Artificial Intelligence) tasks.
7Integrated GitHub marketplace for discovering, installing, and submitting AI (Artificial Intelligence) customization elements.
8Local-first architecture with GitHub sync, ensuring elements are stored locally, work offline, and can be backed up.
9MCP-AQL (Model Context Protocol - Agent Query Language) for structured operations and five semantic endpoints.
10Server-side permission enforcement via Gatekeeper, allowing active Dollhouse elements to shape AI (Artificial Intelligence) actions.

use cases

Who Should Use mcp-server?

mcp-server is designed for a broad audience seeking to customize and control AI (Artificial Intelligence) behavior, from individual developers to large professional teams. Its modular and open-source nature supports diverse applications across various domains.

1**Developers:** For building AI (Artificial Intelligence)-powered applications, orchestrating multi-agent workflows, and leveraging community-powered code review personas and debugging skills.
2**Writers:** For creating and utilizing community-crafted writing styles and templates, refining AI (Artificial Intelligence) prompts, and adapting AI (Artificial Intelligence) language and attitudes.
3**Professionals:** For customizing AI (Artificial Intelligence) behavior, enhancing knowledge by pointing AI (Artificial Intelligence) to documentation, and leveraging industry-built elements for specific tasks.
4**Researchers/Experimenters:** For multi-model experimentation and comparison, advanced prompt engineering and optimization, and testing different AI (Artificial Intelligence) models with consistent interaction management.
5**Community Contributors:** For sharing and monetizing AI (Artificial Intelligence) customization elements through the integrated GitHub marketplace, fostering collaborative development.

pricing

mcp-server Pricing & Plans

DollhouseMCP's mcp-server operates on a freemium model. The core package, @dollhousemcp/mcp-server, is open-source and available under the AGPL-3.0-or-later license, allowing users to deploy and utilize the server without direct cost for its foundational functionalities. Specific details regarding premium tiers or hosted services are not publicly detailed, but the freemium designation suggests potential for advanced features or managed offerings beyond the open-source core.

competitors

mcp-server vs Competitors

DollhouseMCP's mcp-server operates within the ecosystem of AI (Artificial Intelligence) customization and persona management tools, distinguishing itself through its open-source Model Context Protocol (MCP) server architecture and focus on markdown-driven persona definitions. It competes with both integrated AI (Artificial Intelligence) features and dedicated persona generation platforms.

Claude AI (Built-in Styles)↗

It is a native feature within the Claude AI platform, allowing users to directly define and apply custom writing styles and personas to their conversations.

Unlike mcp-server, which acts as an external server for persona management, Claude's built-in styles are integrated directly into the AI assistant itself. It offers similar dynamic persona switching but is limited to the Claude ecosystem and doesn't explicitly mention markdown file input for persona definitions.

Juma (formerly Team-GPT)↗

Juma provides an AI customer persona generator that allows users to build custom prompts and generate detailed personas for various AI models, including Claude and ChatGPT.

Juma focuses on generating and refining personas through custom prompts and offers collaboration features, which is similar to mcp-server's goal of managing AI personas. While it supports multiple AI models like mcp-server, its emphasis is on the creation and refinement of persona prompts rather than a server-based protocol for dynamic switching from markdown files.

Semaphor - Assistant Profiles↗

Semaphor allows users to define and customize how embedded AI assistants behave within specific business contexts and dashboards.

Similar to mcp-server, Semaphor focuses on tailoring AI assistant behavior through profiles. It enables customization and reuse of these profiles across different dashboards, providing dynamic context management, though the method of defining profiles might differ from mcp-server's markdown file approach.

OneRun↗

OneRun is an open-source tool designed to create and deploy 'fake AI users' with distinct personas and goals to automatically test chatbots and find bugs.

While its primary use case is AI testing, OneRun directly involves creating and deploying AI personas with specific behaviors, aligning with mcp-server's core functionality of dynamic AI persona management. It is open-source and local, similar to the spirit of a server-based solution, and focuses on the practical application of personas for interaction.

❓

Frequently Asked Questions

+What is mcp-server?

+Is mcp-server free?

Yes, the core package of mcp-server, `@dollhousemcp/mcp-server`, is open-source and available under the AGPL-3.0-or-later license. The overall business model is freemium, suggesting potential for additional paid features or services.

+What are the main features of mcp-server?

Key features include dynamic AI (Artificial Intelligence) persona management from markdown files, AI (Artificial Intelligence) persona activation and switching for compatible AI (Artificial Intelligence) assistants like Claude (other), API (Application Programming Interface) availability, creation and sharing of modular AI (Artificial Intelligence) personas and skills, multi-agent workflow orchestration, and an integrated GitHub marketplace for elements.

+Who should use mcp-server?

mcp-server is intended for developers building AI (Artificial Intelligence)-powered applications, writers crafting AI (Artificial Intelligence) styles, professionals customizing AI (Artificial Intelligence) behavior, researchers for multi-model experimentation, and community contributors sharing AI (Artificial Intelligence) customization elements.

+How does mcp-server compare to alternatives?

mcp-server distinguishes itself by offering an external, open-source Model Context Protocol (MCP) server for markdown-driven persona management, unlike Claude AI's built-in styles. It focuses on dynamic switching via a server protocol, differing from Juma's prompt-based persona generation or Semaphor's embedded AI (Artificial Intelligence) profiles. While OneRun also uses AI (Artificial Intelligence) personas, mcp-server's primary use case is broader AI (Artificial Intelligence) customization rather than solely chatbot testing.