MCP Prompt Manager

MCP Prompt Manager is an MCP server that provides AI assistants such as Claude Desktop with full access to a local file-based prompt management system. Designed for prompt engineers, developers, and researchers, it enables language models to perform CRUD operations on prompt files stored on the user's local disk. The server exposes an extensive suite of tools for prompt cataloging, including fuzzy search based on Levenshtein and n-gram similarity, tagging, categorization, and favorites management. In addition to basic file retrieval and updates, the server incorporates a template engine supporting variable substitution and conditional syntax, paired with a built-in library of professional templates spanning software development, document drafting, language translation, and research. MCP Prompt Manager also handles prompt lifecycle management through granular version control, allowing users to view change histories, inspect diffs across revisions, and perform rollbacks. Configurable via environment variables for caching, rate limiting, and access permissions, it facilitates secure prompt organization and backup workflows using standard JSON import and export functionality.

Category: AI & LLM Tooling

Tags: file-management, prompt-engineering, prompts, templates

Visit MCP Prompt Manager

How to install and configure MCP Prompt Manager

1. Clone the repository and navigate to its directory: bash git clone https://github.com/Tae4an/mcp-prompt-manager.git cd mcp-prompt-manager 2. Install project dependencies and grant execution permissions: bash npm install chmod +x server.js 3. Open your Claude Desktop configuration file (~/Library/Application Support/Claude/claude_desktop_config.json on macOS or %APPDATA%\Claude\claude_desktop_config.json on Windows). 4. Register the server in the mcpServers object: json { "mcpServers": { "promptManager": { "command": "node", "args": ["absolute_path_to_cloned_repository/server.js"] } } } 5. Restart Claude Desktop.

What you can do with MCP Prompt Manager

• Searching local prompt libraries using fuzzy search algorithms to quickly locate relevant templates based on filenames or text content during LLM conversations. - Rendering modular prompt templates with dynamic variable substitution using the double-curly brace syntax to generate tailored inputs for AI tasks. - Tracking prompt revisions over time by comparing diffs across saved versions and rolling back prompts to earlier states when needed. - Organizing custom prompt collections using metadata tags, categories, and favorites flags to maintain structured repositories across projects. - Exporting and importing prompt databases with complete version histories and metadata in JSON format for backups or sharing across team environments.

Key facts

• https://github.com/Tae4an/mcp-prompt-manager
• AI & LLM Tooling, Notes & Knowledge Management
• file-management, prompt-engineering, prompts, templates

Part of MCP Servers

How do I install MCP Prompt Manager?

To install MCP Prompt Manager, ensure Node.js version 18 or higher is installed. Clone the repository from GitHub, run npm install in the project directory, and grant execution permissions to server.js. Next, open your Claude Desktop configuration file and add promptManager under the mcpServers configuration block pointing to the absolute path of server.js using node. Finally, restart Claude Desktop.

What can MCP Prompt Manager do?

MCP Prompt Manager allows AI clients to list, retrieve, create, update, and delete local prompt text files. It supports advanced fuzzy search across file titles and contents, provides tagging and categorization tools, renders templates with variable substitution, maintains prompt version history with diff comparisons and rollbacks, and exports or imports entire prompt libraries as JSON.

Which MCP clients work with MCP Prompt Manager?

MCP Prompt Manager works with any Model Context Protocol compliant client, including Claude Desktop. It runs locally as a standard input and output process managed by Node.js or can be deployed inside a Docker container using Docker Compose for containerized development environments.

Does MCP Prompt Manager include pre-built prompt templates?

Yes, the server features a built-in template library with twelve templates across five categories, including coding and development, translation and language, document writing, analysis and research, and education and learning. Users can inspect template variables, render templates directly, or generate new local prompt files from them.