MCP Claude Context Continuity

MCP Claude Context Continuity

An MCP server that maintains conversation context for the Claude CLI across multiple invocations using its built-in resume feature. It enables users to save, restore, and branch sessions to manage continuous AI interactions effectively.

Category
Visit Server

README

MCP Claude Context Continuity

npm version License: MIT

An MCP (Model Context Protocol) server that maintains conversation context for Claude CLI across multiple invocations.

Features

  • 🔄 Conversation Continuity: Leverages Claude CLI's --resume feature to maintain conversation context across multiple calls
  • 📝 Session Management: Save, restore, and branch conversations
  • 🖥️ Cross-platform: Supports Windows (via WSL), Linux, and macOS
  • Simple Implementation: All functionality in a single Python file

Requirements

  • Python 3.8+
  • Claude CLI (must be installed in WSL for all platforms)
  • FastMCP (pip install fastmcp)
  • WSL2 for Windows environments
  • Windows: Python must be installed on Windows (not in WSL)

Installation

Option 1: Using npm (Recommended)

npm install -g mcp-claude-context-continuity

After installation, the server will be available as a global command.

Option 2: Manual Installation

  1. Clone the repository:
git clone https://github.com/tethiro/mcp-claude-context-continuity.git
cd mcp-claude-context-continuity
  1. Install dependencies:
pip install -r requirements.txt

Configuration

Gemini CLI (Recommended)

Add the following to your Gemini configuration file:

Using npm global install (Simplest)

All platforms (~/.gemini/settings.json or %USERPROFILE%\.gemini\settings.json):

{
  "mcpServers": {
    "claude-cli-server": {
      "command": "mcp-claude-context-continuity"
    }
  }
}

Manual installation

Windows (%USERPROFILE%\.gemini\settings.json):

{
  "mcpServers": {
    "claude-cli-server": {
      "command": "wsl",
      "args": [
        "-e",
        "python3",
        "/mnt/c/path/to/mcp-claude-context-continuity/src/claude_cli_server.py"
      ]
    }
  }
}

WSL/Linux/macOS (~/.gemini/settings.json):

{
  "mcpServers": {
    "claude-cli-server": {
      "command": "python3",
      "args": [
        "/path/to/mcp-claude-context-continuity/src/claude_cli_server.py"
      ]
    }
  }
}

Claude Desktop

Add similar configuration to your Claude Desktop settings file.

Usage

Basic Usage

  1. Simple conversation:
execute_claude(prompt="Hello, I'm John")
  1. Continue conversation:
execute_claude(prompt="Do you remember my name?")
→ "Yes, you're John"

Session Management

  1. Save current session:
session_id = get_current_session()
  1. Start new session:
reset_session()
  1. Restore saved session:
set_current_session(session_id="saved_id")

Available Tools

Tool Description
execute_claude Execute Claude CLI with conversation continuity
execute_claude_with_context Execute with file context
get_execution_history Get execution history
get_current_session Get current session ID
set_current_session Set session ID
reset_session Reset session
clear_execution_history Clear history
test_claude_cli Test functionality

How Session Management Works

Using Claude CLI's --resume feature, you can return to any point in a conversation using session IDs:

Conversation 1: "I'm Alice" → session_id: AAA
Conversation 2: "I like reading" (--resume AAA) → session_id: BBB
Conversation 3: "My favorite book is..." (--resume BBB) → session_id: CCC

Later:
set_current_session("AAA") → State where only "Alice" is known
set_current_session("BBB") → State where "Alice" and "reading" are known

Troubleshooting

Claude CLI Not Found

Set the CLAUDE_PATH environment variable:

export CLAUDE_PATH=/path/to/claude

Windows Environment Notes

  • WSL2 is required
  • Install Claude CLI inside WSL
  • Use WSL format paths (/mnt/c/...)

License

MIT License


MCP Claude Context Continuity

Claude CLIの会話コンテキストを保持するMCP(Model Context Protocol)サーバーです。

特徴

  • 🔄 会話の継続性: Claude CLIの--resume機能を活用し、複数の呼び出し間で会話コンテキストを保持
  • 📝 セッション管理: 会話の保存、復元、分岐が可能
  • 🖥️ クロスプラットフォーム: Windows(WSL経由)、Linux、macOS対応
  • シンプルな実装: 単一のPythonファイルで全機能を提供

必要条件

  • Python 3.8以上
  • Claude CLI(WSL内にインストール必須)
  • FastMCP(pip install fastmcp
  • Windows環境の場合はWSL2
  • Windows: PythonはWindows側にインストール(WSL内ではなく)

インストール

方法1: npmを使用(推奨)

npm install -g mcp-claude-context-continuity

インストール後、グローバルコマンドとして利用可能になります。

方法2: 手動インストール

  1. リポジトリをクローン:
git clone https://github.com/tethiro/mcp-claude-context-continuity.git
cd mcp-claude-context-continuity
  1. 依存関係をインストール:
pip install -r requirements.txt

設定

Gemini CLI(推奨)

Geminiの設定ファイルに以下を追加:

npmグローバルインストールを使用(最も簡単)

全環境共通 (~/.gemini/settings.json または %USERPROFILE%\.gemini\settings.json):

{
  "mcpServers": {
    "claude-cli-server": {
      "command": "mcp-claude-context-continuity"
    }
  }
}

手動インストールの場合

Windows環境 (%USERPROFILE%\.gemini\settings.json):

{
  "mcpServers": {
    "claude-cli-server": {
      "command": "wsl",
      "args": [
        "-e",
        "python3",
        "/mnt/c/path/to/mcp-claude-context-continuity/src/claude_cli_server.py"
      ]
    }
  }
}

WSL/Linux/macOS (~/.gemini/settings.json):

{
  "mcpServers": {
    "claude-cli-server": {
      "command": "python3",
      "args": [
        "/path/to/mcp-claude-context-continuity/src/claude_cli_server.py"
      ]
    }
  }
}

Claude Desktop

Claude Desktopの設定ファイルに同様の設定を追加します。

使用方法

基本的な使い方

  1. 通常の会話:
execute_claude(prompt="こんにちは、私は山田です")
  1. 会話の継続:
execute_claude(prompt="私の名前を覚えていますか?")
→ "はい、山田さんですね"

セッション管理

  1. 現在のセッションを保存:
session_id = get_current_session()
  1. 新しいセッションを開始:
reset_session()
  1. 保存したセッションに戻る:
set_current_session(session_id="保存したID")

利用可能なツール

ツール 説明
execute_claude Claude CLIを実行(会話継続)
execute_claude_with_context ファイルコンテキスト付きで実行
get_execution_history 実行履歴を取得
get_current_session 現在のセッションIDを取得
set_current_session セッションIDを設定
reset_session セッションをリセット
clear_execution_history 履歴をクリア
test_claude_cli 動作確認

セッション管理の仕組み

Claude CLIの--resume機能を活用し、セッションIDによって会話の任意の時点に戻ることができます:

会話1「私は太郎です」 → session_id: AAA
会話2「趣味は読書です」(--resume AAA) → session_id: BBB
会話3「好きな本は...」(--resume BBB) → session_id: CCC

後から:
set_current_session("AAA") → "太郎"のみ知っている状態
set_current_session("BBB") → "太郎"と"読書"を知っている状態

トラブルシューティング

Claude CLIが見つからない場合

環境変数CLAUDE_PATHを設定:

export CLAUDE_PATH=/path/to/claude

Windows環境での注意点

  • WSL2が必要です
  • Claude CLIはWSL内にインストールしてください
  • ファイルパスはWSL形式(/mnt/c/...)で指定

ライセンス

MIT License

Recommended Servers

playwright-mcp

playwright-mcp

A Model Context Protocol server that enables LLMs to interact with web pages through structured accessibility snapshots without requiring vision models or screenshots.

Official
Featured
TypeScript
Magic Component Platform (MCP)

Magic Component Platform (MCP)

An AI-powered tool that generates modern UI components from natural language descriptions, integrating with popular IDEs to streamline UI development workflow.

Official
Featured
Local
TypeScript
Audiense Insights MCP Server

Audiense Insights MCP Server

Enables interaction with Audiense Insights accounts via the Model Context Protocol, facilitating the extraction and analysis of marketing insights and audience data including demographics, behavior, and influencer engagement.

Official
Featured
Local
TypeScript
VeyraX MCP

VeyraX MCP

Single MCP tool to connect all your favorite tools: Gmail, Calendar and 40 more.

Official
Featured
Local
graphlit-mcp-server

graphlit-mcp-server

The Model Context Protocol (MCP) Server enables integration between MCP clients and the Graphlit service. Ingest anything from Slack to Gmail to podcast feeds, in addition to web crawling, into a Graphlit project - and then retrieve relevant contents from the MCP client.

Official
Featured
TypeScript
Kagi MCP Server

Kagi MCP Server

An MCP server that integrates Kagi search capabilities with Claude AI, enabling Claude to perform real-time web searches when answering questions that require up-to-date information.

Official
Featured
Python
E2B

E2B

Using MCP to run code via e2b.

Official
Featured
Neon Database

Neon Database

MCP server for interacting with Neon Management API and databases

Official
Featured
Exa Search

Exa Search

A Model Context Protocol (MCP) server lets AI assistants like Claude use the Exa AI Search API for web searches. This setup allows AI models to get real-time web information in a safe and controlled way.

Official
Featured
Qdrant Server

Qdrant Server

This repository is an example of how to create a MCP server for Qdrant, a vector search engine.

Official
Featured