MCP Server Validator
Validate your Model Context Protocol manifest — required fields, tool schemas, description quality
initialize) below.How to use
- Paste your MCP manifest JSON
- Review errors and warnings
- Iterate until all checks pass
Use cases
- MCP server development
- Claude tool-use QA
- Agent debugging
About the MCP Server Validator
The Model Context Protocol (MCP) is Anthropic's 2024 open standard for connecting LLMs to tools and data sources. Its ecosystem is growing fast — Claude Desktop, Zed, Continue, and more. Getting the manifest right matters because the descriptions you write directly shape whether the LLM calls your tools correctly. This validator catches structural and quality issues before you ship.
Features
- Validates required top-level fields (name, version, protocolVersion, capabilities)
- Checks every tool has name, description, inputSchema
- Warns on vague descriptions (LLMs choose tools by description — bad ones = wrong tool calls)
- Validates inputSchema is proper JSON Schema with type: object
- Flags missing property descriptions
- Resource and prompt validation
How it works
- Paste your MCP server manifest JSON (the response from initialize or your server.json).
- Review errors (must-fix) and warnings (should-fix).
- Iterate until all checks pass.
Frequently asked questions
What is MCP?
+
Model Context Protocol, introduced by Anthropic in Nov 2024. It's an open standard for connecting LLMs to data sources and tools — think USB-C for AI. Claude Desktop, Zed, and several other clients support it.
Why does description quality matter so much?
+
The LLM chooses which tool to call based on the description. Vague descriptions like 'do stuff' lead to wrong tool selection. Clear ones like 'Fetches the current weather for a given city; use when the user asks about weather, temperature, or conditions' dramatically improve accuracy.
Is this the MCP spec as of 2024-11-05 or 2025-03-26?
+
Validation covers both protocol versions. We warn if you're on the older 2024-11-05 version — upgrading is usually easy and adds useful features.
Can I test the server runtime, not just the manifest?
+
This validates the static manifest only. Runtime testing (initialize → list_tools → call) requires a running server — use the MCP Inspector from modelcontextprotocol.io for that.
Does it validate JSON-RPC message schemas?
+
Not yet. Manifest validation only.