JSON

What it means

JSON is the most common interchange format for structured model output: field names, typed values, and arrays the product can validate and render predictably.

Why designers should care

JSON-backed flows need visible schemas, field-level errors, and preview states so users trust machine-readable output before it hits production systems.

Example

A “Generate API spec” tool returns JSON with endpoints and params; the UI renders an editable table, highlights invalid types, and exports only after validation passes.

Common mistakes

• • Showing raw JSON blobs with no field labels or human-readable preview.
• • Assuming the model always returns valid JSON without retry or repair UX.
• • Hiding structure from non-technical users who still need to approve content.

Related terms

• Schema: A schema is the contract between prompt, model output, and interface: which fields exist, what they mean, and what “valid” looks like.
• Structured Output: The product constrains generation to fields, types, and enums (often via schema validation) so downstream code and design systems can render reliably.
• Inference: Each user request triggers inference: tokens in, tokens out, optionally interleaved with tool calls and retrieval.
• Tool Use: The model emits structured calls (name + arguments) that your app executes, then feeds results back into the conversation for the next step.
• Markdown (MD): Markdown is structured text without heavy syntax: `#` for headings, `-` for bullets, fenced blocks for code, and links in `[label](url)` form.