Anthropic Tool Input Builder
Build Claude tool definitions with `input_schema` from sample JSON input
Outputs a Claude tool definition entry with `name`, `description`, and `input_schema`.
Sample JSON
Generated Output
The schema is generated entirely in your browser from the sample you paste here. No provider APIs or server-side processing are involved.
Build Claude tool definitions from a real input sample
Claude tool calling expects a tool definition with name, description, and input_schema. Hand-writing JSON Schema for every tool is slow and error-prone. Pasting a real input sample produces a complete tool definition you can drop directly into the tools array of your Anthropic request.
Use the builder when you need to
Define a new Claude tool quickly
Generate the full tool object — name, description, and input_schema — from a sample arguments JSON.
Migrate tools from another provider
Take an OpenAI function definition or MCP tool and rebuild it as a Claude-shaped input_schema.
Tighten an existing tool's schema
Regenerate from a more representative sample so optional fields and enums match production reality.
How to build a Claude tool definition
- 1
Paste a sample of the tool's input arguments as JSON.
- 2
Set the tool name and description for Claude.
- 3
Copy the generated tool object directly into your Anthropic request's tools array.
Keep going
Build for other providers
Use the broader builder for OpenAI Structured Outputs and MCP definitions.
Validate MCP definitions
Run validation when you're publishing the same tool through an MCP server.
Generate plain JSON Schema
Pick this when you need a schema for non-Anthropic consumers.
Validate Claude's output
Check that Claude's tool calls match the schema you defined.
Common Claude tool workflows
Generate the input_schema, drop it in the tools array, and start handling tool calls.
Regenerate the schema from a new sample when you change the tool's expected input.
Reuse the same input shape and emit it in the format Claude expects.
Related Tools
Structured Output Schema Builder
Generate OpenAI, Anthropic, MCP, or plain JSON Schema wrappers from sample output
MCP Tool Schema Validator
Validate MCP tool definitions and verify `inputSchema` before deployment
JSON Schema Generator
Generate JSON Schema from sample JSON for apps, APIs, and AI workflows
LLM JSON Output Validator
Repair and validate model JSON output against a schema
Frequently Asked Questions
It outputs a Claude tool definition object with `name`, `description`, and `input_schema`, ready to paste into the `tools` array of your Anthropic request.
Anthropic tool inputs are usually named parameters, so the schema should have an object root with properties rather than a raw string or array at the top level.
Yes. This page only generates the JSON definition. You still send that JSON from your own application when you call Anthropic, but the building step itself is entirely local.
Yes. This tool is focused specifically on Anthropic's `input_schema` wrapper. The broader builder also supports plain JSON Schema, OpenAI Structured Outputs, and MCP definitions.