Project Description MCP Server

@pizzaeueu

リポジトリ

基本情報

カテゴリ

AI とエージェント

トランスポート

stdio

作者

pizzaeueu

Project Description MCP Server について

The proxy enables integration between OpenAI chat models and local MCP servers via Function Calling, allowing custom PII checks for sensitive data verification

Project Description MCP Server の設定

標準の設定はありません

このサーバーの README には解析可能な MCP 設定ブロックが含まれていません。インストール手順はリポジトリをご確認ください。

リポジトリ

Project Description MCP Server のツール

ツール一覧はまだありません

この掲載には表示できるツール一覧がまだありません。サーバーにツールがないという意味ではありません。対応機能はプロジェクトのドキュメントをご確認ください。

ツールは README から自動的に抽出されます。メンテナーは ## Tools という見出しの下に記載することで、このタブに反映できます。

概要

What is Project Description?

Project Description is a proxy that integrates OpenAI chat models with local MCP servers via function calling, adding custom PII checks for sensitive data verification. It is designed for users who need to control whether sensitive information retrieved from MCP tools is shared with the language model.

How to use Project Description?

Set your OpenAI API key in the OPEN_AI_API_KEY variable and an absolute path to the shared folder in the SHARED_DIRECTORY_ABSOLUTE_PATH variable in the Docker Compose file, then run docker compose up. Open http://localhost:4173/ and start chatting with the model. By default, it works with the filesystem MCP server (pre‑installed as a Docker image), but you can configure other MCP servers by editing data/McpConfig.conf.

Key features of Project Description

Enables OpenAI models to call tools on local MCP servers.
Runs regex‑based PII checks on data retrieved from MCP servers.
Asks users for permission before sharing sensitive data with the LLM.
Stateful with in‑memory persistence (workflows lost on crash/restart).
No authentication mechanism, no metrics, no strict SLAs.
PII detection supports English language only.

Use cases of Project Description

Safely use MCP file‑system tools with OpenAI without leaking sensitive content.
Review and approve or deny sharing of detected PII before it reaches the LLM.
Prototype integrations between OpenAI and custom MCP servers in a local environment.

よくある質問

What does Project Description do differently from a direct MCP integration?

It adds a PII verification layer between MCP tool results and the LLM, prompting the user to allow or block sharing of sensitive data before the conversation continues.

What dependencies or runtime are required?

OpenAI chat models with tool‑calling support, a Docker environment, and an MCP server (e.g., the filesystem server installed as a Docker image).

Where does user data live?

User data is stored in‑memory; it is lost on crash or restart, and the app is not horizontally scalable.

What are the known limits of Project Description?

PII detection only works for English. If the data retrieved from MCP servers exceeds the LLM’s context window, the app returns an error and stops the dialog. Adding a new MCP server requires a restart.

What transport or authentication does it use?

The proxy does not implement any authentication mechanism. It communicates with MCP servers via the MCP Java SDK and with OpenAI via the OpenAI Java SDK using function calling.

「AI とエージェント」の他のコンテンツ

Gemina

Gemina Tech Ltd.

Extract, search and tag any document — invoices, receipts, contracts, forms, or your own templates. 13 tools, one sign-in (OAuth 2.1) or API key, free FileTag tier, data residency in the EU, US, Israel or Asia.

HeyLead: LinkedIn Outreach & Lead Generation MCP Server

D4umak

HeyLead: AI agent for LinkedIn outreach. Finds the right people, writes in your voice, follows up and handles replies, from Claude Code, Cursor, any MCP client or a web dashboard.

fhirHydrant

faulkj

Open-source Node.js FHIR MCP server with SMART Backend Services, metadata-aware search/CRUD tools, compact responses, FHIRPath filtering, safe pagination, audit events, and terminology lookup.

Legion MCP