Project Memory MCP
@PyneSys
About Project Memory MCP
An MCP server that enables AI agents to persistently store and retrieve project information from a memory file.
Config
Add this server to your MCP-compatible client using the configuration below.
{
"mcpServers": {
"project-mem-mcp": {
"command": "python",
"args": [
"-m",
"venv",
"venv"
]
}
}
}Tools
No tools detected
We auto-extract tools from the README. The maintainer can list them under a ## Tools heading to populate this section.
Overview
What is Project Memory MCP?
An MCP server that stores and retrieves project information in Markdown files, allowing AI agents (like Claude) to maintain persistent memory about projects between conversations. The memory is stored in a MEMORY.md file in each project directory.
How to use Project Memory MCP?
Install via uvx or from source. Configure the MCP client (e.g., Claude Desktop) with a JSON entry pointing to the server and specifying allowed directories. The client starts the server automatically. Use the three provided tools: get_project_memory to read memory at the start of a conversation, set_project_memory to overwrite it, and update_project_memory for efficient patching.
Key features of Project Memory MCP
- Stores project information in Markdown format
- Retrieves memory at the beginning of conversations
- Updates memory using block-based SEARCH/REPLACE patches
- Validates project paths against allowed directories
- Restricts all file operations to allowed directories
Use cases of Project Memory MCP
- AI agents maintaining persistent memory about a project across multiple chats
- Claude Desktop users linking project instructions to a memory file
- Updating project information incrementally during a conversation
- Falling back to full memory replacement if a patch fails
FAQ from Project Memory MCP
What does the --allowed-dir argument do?
It specifies the directories the server has access to. You can use it multiple times; if omitted, the server only accesses the user's home directory.
How is the server started?
The MCP client (e.g., Claude Desktop) starts the server based on your configuration. You do not need to start it manually.
What tools does Project Memory MCP provide?
Three tools: get_project_memory (reads memory), set_project_memory (overwrites memory), and update_project_memory (applies block-based patches).
What format is the project memory stored in?
The memory is stored in a MEMORY.md file inside the project directory, using Markdown formatting.
Are there security considerations?
Yes. Memory files should never contain sensitive information. All file operations are restricted to allowed directories, and project paths are validated against those directories.
Frequently asked questions
What does the `--allowed-dir` argument do?
It specifies the directories the server has access to. You can use it multiple times; if omitted, the server only accesses the user's home directory.
How is the server started?
The MCP client (e.g., Claude Desktop) starts the server based on your configuration. You do not need to start it manually.
What tools does Project Memory MCP provide?
Three tools: `get_project_memory` (reads memory), `set_project_memory` (overwrites memory), and `update_project_memory` (applies block-based patches).
What format is the project memory stored in?
The memory is stored in a `MEMORY.md` file inside the project directory, using Markdown formatting.
Are there security considerations?
Yes. Memory files should never contain sensitive information. All file operations are restricted to allowed directories, and project paths are validated against those directories.
Basic information
More Memory & Knowledge MCP servers
The Cloud
TekaThe Cloud is the AI-native workspace that gives every AI — Claude, ChatGPT, Gemini, and other Frontier & Free or Local models — shared memory and real agents. Not cloud computing: thecloud.so is one workspace where your
Memory
modelcontextprotocolModel Context Protocol Servers
Docs MCP Server
araboldGrounded Docs MCP Server: Open-Source Alternative to Context7, Nia, and Ref.Tools
OpenLore
aakarimServe your docs to AI agents over SSH and MCP. Agents use ls, cat, grep, find, and pipes against a virtual filesystem backed by your documentation. Identity-scoped access, governed writing with human approval, and portab
Solomd
zhitongblogA markdown editor — and the bridge to your LLM. Local-first, MIT, ~15 MB. Bundled MCP server lets Claude Code / Codex / Cursor drive your vault directly. 14 AI providers BYOK.
Comments