Blog

We're shifting our writing to jsontotypescript.com. Here's why.

Use generated TypeScript interfaces from real JSON to eliminate any types across a legacy codebase.

Compare two API response versions by diffing their schemas instead of wading through raw JSON.

Use schema snapshots to detect unexpected API response changes in your CI pipeline.

Use compact schemas in your prompts to tell LLMs exactly what JSON structure to return.

JSON Schema is verbose for LLM prompts. A compact schema conveys the same structure in a fraction of the tokens.

Use generated schemas to document API endpoints faster than writing examples or specs by hand.

When JSON array items differ, schema generation merges them into one shape with optional fields.

Generate TypeScript interfaces from your JSON config files to catch typos and missing fields at compile time.

How to generate TypeScript interfaces from JSON responses when the API has no type definitions.

Convert unknown API responses to schemas to understand their structure in seconds, not minutes.

Turn sample JSON into function-calling parameter schemas instead of writing them by hand.

Pipe API responses from curl into the jsontoschema CLI to get TypeScript interfaces in seconds.

When to use a generated schema and when to reach for Zod for validating JSON at runtime.

Raw JSON wastes tokens and confuses models. A compact schema gives the LLM the structure it needs in a fraction of the size.