LLM JSON Schema Builder

Structured output forces LLMs to return data in a specific JSON format rather than free-form text. OpenAI's JSON mode and Anthropic's tool use feature guarantee the response is valid JSON matching a defined schema. This is essential for building reliable pipelines, data extraction systems, and agent tools.

JSON Schema is a vocabulary that allows you to annotate and validate JSON documents. It defines the structure of JSON data: field names, types (string, number, boolean, array, object), required fields, enums, and validation constraints. LLM APIs use JSON Schema to define the expected output format.

Pass the generated JSON Schema to OpenAI's response_format parameter: `response_format: { type: 'json_schema', json_schema: { name: 'YourSchema', schema: <your-schema>, strict: true } }`. For older JSON mode, just set `response_format: { type: 'json_object' }` and describe the format in your prompt.

For Claude, use the tools feature: define a tool with your JSON Schema as the input_schema. Prompt Claude to use that tool. Claude will return a tool_use block with the structured data. This is the most reliable way to get structured output from Claude.

Strict mode (available on GPT-4o and later) guarantees 100% adherence to the schema. All object properties must be listed in 'required', additionalProperties must be false, and nested schemas must also be strict. It eliminates hallucinated fields but requires a more complete schema definition.