hivecode
Real-time analytics dashboard for Claude Code, tracking sessions, tool usage, file changes, and token costs via an MCP server.
README
HiveCode
Real-time analytics dashboard for Claude Code
Track sessions, tool usage, file changes, token costs, and more — all from a real-time local dashboard. Built as an MCP server that plugs directly into Claude Code.
Features
- Real-time Dashboard — Live WebSocket updates as Claude Code works
- Session Tracking — Monitor every Claude Code session with full detail
- Tool Analytics — See which tools are used most, average durations, success rates
- File Change Log — Track every file created, edited, or deleted
- Cost Tracking — Token usage and estimated API costs with daily breakdowns
- Canvas Charts — Visual bar and line charts for activity trends
- Activity Feed — Live stream of all events as they happen
- MCP Server — Plugs directly into Claude Code via Model Context Protocol
- Update Notifications — Get notified when a new version is available
- Mobile Responsive — Works on phone and tablet
- 100% Local — All data stored in SQLite on your machine
Quick Start
# Clone the repo
git clone https://github.com/ZedoxDevelopment/hivecode.git
cd hivecode
# Install dependencies
npm install
# Start the dashboard
npm start
Open http://localhost:3800 in your browser.
Seed Demo Data
Want to see the dashboard in action before connecting Claude Code?
node src/seed-demo.js
Connect to Claude Code
Add HiveCode as an MCP server in your Claude Code settings:
~/.claude/settings.json
{
"mcpServers": {
"hivecode": {
"command": "node",
"args": ["/path/to/hivecode/src/mcp-server.js"],
"env": {}
}
}
}
Or run the MCP server standalone:
npm run mcp
MCP Tools
| Tool | Description |
|---|---|
hivecode_start_session |
Start tracking a new Claude Code session |
hivecode_log_tool |
Log a tool call with duration and success status |
hivecode_log_file |
Log a file change (create/edit/delete) |
hivecode_log_tokens |
Log token usage for cost tracking |
hivecode_end_session |
End the current session with a summary |
hivecode_stats |
Get dashboard statistics |
API Endpoints
| Method | Endpoint | Description |
|---|---|---|
| GET | /api/stats |
Dashboard overview stats |
| GET | /api/sessions |
List all sessions |
| GET | /api/sessions/:id |
Session detail with events |
| GET | /api/tools |
Tool usage analytics |
| GET | /api/files |
Recent file changes |
| GET | /api/daily |
Daily stats for charts |
| POST | /api/event |
Log events via HTTP |
| GET | /api/version |
Current version info |
| GET | /api/version/check |
Check for updates |
| GET | /api/health |
Health check |
Dashboard Pages
- Overview — Stats grid, 14-day activity chart, recent sessions, top tools
- Sessions — Searchable/filterable session list with drill-down detail view
- Tools — Usage distribution chart, per-tool bars, performance table
- Files — Searchable file change log across all sessions
- Costs — Cost/token trend charts with daily breakdown table
- Activity Feed — Real-time event stream via WebSocket
- Settings — MCP config, data export, seed/clear data
Configuration
| Environment Variable | Default | Description |
|---|---|---|
HIVECODE_PORT |
3800 |
Dashboard server port |
Tech Stack
- Backend: Express.js, WebSocket (ws), better-sqlite3
- Frontend: Vanilla JS, Canvas charts, CSS custom properties
- Protocol: MCP (Model Context Protocol) over JSON-RPC stdio
- Database: SQLite with WAL mode
Project Structure
hivecode/
src/
server.js # Express + WebSocket server
db.js # SQLite schema + prepared statements
mcp-server.js # MCP protocol handler + tools
seed-demo.js # Demo data generator
public/
index.html # Dashboard shell
app.js # Frontend application
style.css # Dark theme styles
data/ # SQLite database (auto-created)
version.json # Version + changelog for update checks
package.json
License
MIT — Zedox Development
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。
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。
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。
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。
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。
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。
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。
mcp-server-qdrant
这个仓库展示了如何为向量搜索引擎 Qdrant 创建一个 MCP (Managed Control Plane) 服务器的示例。
e2b-mcp-server
使用 MCP 通过 e2b 运行代码。