Tool Call Schema Validator Connector for AI agents.

It acts as a buffer that checks every argument an agent tries to send against your specific rules. If the agent tries to send invalid data, the MCP catches it before the call ever reaches your API.

Yes. It is designed to perform recursive checks, meaning it can validate complex, multi-layered objects and arrays just as easily as simple key-value pairs.

It uses standard JSON Schema, which is much more robust and easier to maintain than writing custom validation logic for every single capability your agent uses.

It provides exact paths to errors. Instead of a generic error message, you get the specific location in your data, like args.settings.theme, so you know exactly what to fix.

Yes, it is built on the open standard, so you can connect it to any MCP-compatible client like Claude, Cursor, or Windsurf to add validation to your agentic loops.

It validates that LLM-generated arguments strictly match a provided JSON Schema, including types, required fields, and logical constraints like min/max values.

Errors are reported using precise path notation, such as args.items[0].id, so you know exactly which part of the input failed validation.

Yes, the validator uses recursive traversal to ensure every level of a nested object or array is checked against the schema.