
Context Continuation MCP Server
Provides intelligent context management for AI development sessions, allowing users to track token usage, manage conversation context, and seamlessly restore context when reaching token limits.
README
Context Continuation MCP Server
An MCP (Model Context Protocol) server that provides intelligent context management for AI development sessions. Never lose context when hitting token limits again!
Features
- Automatic Context Tracking: Monitor token usage and conversation flow
- Intelligent Session Breaks: Get notified before hitting context limits
- Seamless Restoration: Generate context restoration prompts for new sessions
- Project Management: Track milestones, decisions, and progress across sessions
- File-Based Storage: Human-readable markdown files that work with git
Quick Start
Installation
npm install -g context-continue-mcp
Usage with Claude Desktop
- Add to your Claude Desktop configuration:
{
"mcpServers": {
"context-continue": {
"command": "context-mcp",
"args": ["--project", "/path/to/your/project"]
}
}
}
-
Restart Claude Desktop
-
Start using context tools in your conversations:
context_start_session
- Begin tracking a new sessioncontext_track_message
- Track important messagescontext_get_status
- Check token usagecontext_restore_session
- Generate restoration prompt
Tools Available
Session Management
context_start_session
- Start tracking a new context sessioncontext_end_session
- End current session with summarycontext_get_status
- Get current session and token usage info
Context Tracking
context_track_message
- Add message to session trackingcontext_track_progress
- Update project progresscontext_add_milestone
- Add project milestone
Restoration
context_restore_session
- Generate context restoration promptcontext_get_project_summary
- Get full project overview
How It Works
- Start a Session: Initialize context tracking for your project
- Track Progress: Important messages and decisions are automatically logged
- Monitor Usage: Get warnings when approaching token limits
- Seamless Continuation: Generate restoration prompts for new sessions
File Structure
The server creates a .context
directory in your project:
your-project/
├── .context/
│ ├── config.json
│ ├── project_summary.md
│ ├── sessions/
│ │ ├── session_001_2025-05-31.md
│ │ └── session_002_2025-06-01.md
│ ├── progress/
│ │ ├── milestones.md
│ │ └── decisions.md
│ └── artifacts/
└── your-code/
Quality Assurance
This project maintains high code quality through:
- 🧪 Comprehensive Testing: 43+ unit tests with 95%+ coverage
- 🔄 Continuous Integration: Automated testing on Node.js 18.x, 20.x, 21.x
- 🌍 Cross-Platform: Tested on Ubuntu, Windows, and macOS
- 📊 Code Coverage: Real-time coverage tracking with Codecov
- 🏗️ Build Verification: Automated build and CLI functionality testing
- 📦 Package Validation: Pre-publish testing and compatibility checks
Running Tests
# Run all tests
npm test
# Run tests with coverage
npm run test:coverage
# Run specific test suites
npm test token-counter
npm test session-tracker
npm test context-manager
# Watch mode for development
npm run test:watch
Development
git clone https://github.com/core3-coder/context-continue-mcp
cd context-continue-mcp
npm install
npm run build
npm start
License
MIT - see LICENSE file for details
推荐服务器

Baidu Map
百度地图核心API现已全面兼容MCP协议,是国内首家兼容MCP协议的地图服务商。
Playwright MCP Server
一个模型上下文协议服务器,它使大型语言模型能够通过结构化的可访问性快照与网页进行交互,而无需视觉模型或屏幕截图。
Magic Component Platform (MCP)
一个由人工智能驱动的工具,可以从自然语言描述生成现代化的用户界面组件,并与流行的集成开发环境(IDE)集成,从而简化用户界面开发流程。
Audiense Insights MCP Server
通过模型上下文协议启用与 Audiense Insights 账户的交互,从而促进营销洞察和受众数据的提取和分析,包括人口统计信息、行为和影响者互动。

VeyraX
一个单一的 MCP 工具,连接你所有喜爱的工具:Gmail、日历以及其他 40 多个工具。
graphlit-mcp-server
模型上下文协议 (MCP) 服务器实现了 MCP 客户端与 Graphlit 服务之间的集成。 除了网络爬取之外,还可以将任何内容(从 Slack 到 Gmail 再到播客订阅源)导入到 Graphlit 项目中,然后从 MCP 客户端检索相关内容。
Kagi MCP Server
一个 MCP 服务器,集成了 Kagi 搜索功能和 Claude AI,使 Claude 能够在回答需要最新信息的问题时执行实时网络搜索。

e2b-mcp-server
使用 MCP 通过 e2b 运行代码。
Neon MCP Server
用于与 Neon 管理 API 和数据库交互的 MCP 服务器
Exa MCP Server
模型上下文协议(MCP)服务器允许像 Claude 这样的 AI 助手使用 Exa AI 搜索 API 进行网络搜索。这种设置允许 AI 模型以安全和受控的方式获取实时的网络信息。