TMetric MCP Server
Integrates with TMetric for time tracking, enabling project listing, timer start/stop, and deletion of entries, with GitLab issue support.
README
TMetric MCP Server
Minimal Model Context Protocol server for TMetric time tracking integration.
Features
- List TMetric projects
- Start/stop timers
- Check current timer status
- GitLab issue integration
- Delete time entries
Installation
npm install
npm run build
Configuration
Set your TMetric API token:
export TMETRIC_API_TOKEN="your_token_here"
Usage with Claude Code
Add the TMetric MCP server using the claude mcp add command. You can install at different scopes:
User Scope (Global - All Projects)
Available across all projects for your user:
One-liner:
claude mcp add --scope user tmetric-mcp --env TMETRIC_API_TOKEN=your_token_here -- node /path/to/tmetric-mcp-server/build/index.js
Interactive:
claude mcp add tmetric-mcp --scope user
When prompted, configure:
- Command:
node - Args:
/path/to/tmetric-mcp-server/build/index.js - Environment variables:
TMETRIC_API_TOKEN=your_token_here
Project Scope (Specific Project)
Available only in the current project directory:
One-liner:
claude mcp add --scope project tmetric-mcp --env TMETRIC_API_TOKEN=your_token_here -- node /path/to/tmetric-mcp-server/build/index.js
Interactive:
claude mcp add tmetric-mcp --scope project
When prompted, configure with the same settings as above.
Local Scope (Current Directory)
Available only in the current working directory:
One-liner:
claude mcp add --scope local tmetric-mcp --env TMETRIC_API_TOKEN=your_token_here -- node /path/to/tmetric-mcp-server/build/index.js
Interactive:
claude mcp add tmetric-mcp --scope local
When prompted, configure with the same settings as above.
Alternative: Using npx (from GitHub)
You can run directly from GitHub with npx (no clone or build needed):
One-liner:
claude mcp add --scope user tmetric-mcp --env TMETRIC_API_TOKEN=your_token_here -- npx -y github:guglielmo/tmetric-minimal-mcp
Interactive:
claude mcp add tmetric-mcp --scope user
When prompted, configure:
- Command:
npx - Args:
-y github:guglielmo/tmetric-minimal-mcp - Environment variables:
TMETRIC_API_TOKEN=your_token_here
Alternative: Using npx (from local path)
If you've cloned the repository locally:
One-liner:
claude mcp add --scope user tmetric-mcp --env TMETRIC_API_TOKEN=your_token_here -- npx -y /path/to/tmetric-mcp-server
Interactive:
claude mcp add tmetric-mcp --scope user
When prompted, configure:
- Command:
npx - Args:
-y /path/to/tmetric-mcp-server - Environment variables:
TMETRIC_API_TOKEN=your_token_here
Available Tools
list_tmetric_projects()
Get list of available projects.
get_current_timer()
Check if a timer is running.
start_timer(project_id, task_name, task_url?)
Start tracking time on a project/task.
stop_timer()
Stop current timer and return time spent.
delete_time_entry(entry_id?)
Delete a time entry.
Development
# Watch mode
npm run watch
# Run directly with tsx
npm run dev
Testing
Running Tests
# Run all tests once
npm test
# Run tests in watch mode (re-runs on file changes)
npm run test:watch
# Run tests with coverage report
npm run test:coverage
# Run tests with interactive UI
npm run test:ui
Test Coverage
The project has comprehensive test coverage with:
- 100% statement coverage across all modules
- 97%+ branch coverage for edge cases
- Unit tests for all utility functions (
utils.ts) - Full integration tests for TMetric API client (
tmetric-client.ts) - Mocked HTTP requests using
nockfor reliable testing
See TESTING.md for detailed information about the testing strategy and how to write new tests.
Manual Testing
Test the MCP server with Claude Code by starting a conversation and using commands like:
- "List my TMetric projects"
- "Start timer on project 12345 for Issue #123: Fix bug"
- "What am I working on?"
- "Stop the timer"
Troubleshooting
"TMETRIC_API_TOKEN is required"
Make sure you've set the environment variable with your API token.
"Failed to initialize TMetric client"
Check that your API token is valid and you have network access to TMetric.
"Timer already running"
This is expected behavior. Stop the current timer before starting a new one.
Used By
This MCP server is a dependency for the following Claude Code skills:
- issue-time-tracking - Automates synchronized issue status tracking and time logging across GitLab/GitHub and TMetric
License
MIT
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。