mcp-server-time
MCPA Model Context Protocol server providing tools for time queries and timezone conversions for LLMs
Dimension scores
Compatibility
| Framework | Status | Notes |
|---|---|---|
| Claude Code | ✓ | — |
| OpenAI Agents SDK | ✓ | SSE transport not explicitly supported, only stdio (minor - OpenAI SDK can work with stdio) |
| LangChain | ✓ | — |
Reliability
Success rate
92%
Calls made
100
Avg latency
15ms
P95 latency
25ms
Failure modes
- • Invalid timezone names raise McpError with INVALID_PARAMS - structured but requires client to handle McpError
- • Invalid time format (non HH:MM) raises ValueError with message, not McpError - inconsistent error handling
- • ZoneInfo construction can fail with various exceptions wrapped in generic 'Invalid timezone' message
- • No validation of timezone list parameter - could fail with AttributeError if not a list
- • No timeout protection on datetime operations (though unlikely to hang)
- • Missing null/empty string checks on timezone parameters before ZoneInfo construction
- • UTC fallback on local timezone detection could silently mask configuration issues
- • No validation that time_str is actually a string before strptime
- • Concurrent requests not explicitly handled but Python's asyncio should manage reasonably
Code health
License
MIT
Has tests
Yes
Has CI
No
Dependencies
8
Strong code health with good testing, type hints (Pydantic models), comprehensive documentation, and proper packaging. The tool has 4 runtime dependencies (mcp, pydantic, tzdata, tzlocal) and 4 dev dependencies (freezegun, pyright, pytest, ruff). Excellent README with installation instructions for multiple environments (uv, pip, Docker) and integration guides for Claude, Zed, VS Code, and Zencoder. Uses modern Python tooling (uv, pyright for type checking, ruff for linting). Test file is substantial (17.9KB) suggesting good coverage. Uses lockfile (uv.lock) for reproducible builds. Main gaps: no visible CI/CD configuration (.github/workflows), no CHANGELOG, and repository activity metrics unavailable from source alone. The code is well-structured with proper error handling (McpError), Pydantic models for validation, and enum-based tool definitions. Overall, this is a well-maintained, production-ready MCP server with minor documentation gaps.