JSON Schema Inference

Overview

You send one or more example JSON documents and we read their structure to write a conservative JSON Schema in the 2020-12 dialect. A field counts as required when it appears in every example, and a field that holds different types across examples gets a union type plus a warning so you can decide what you actually meant. We stay deliberately plain: no conditionals, no oneOf, no $ref. Use the result as a first draft to review, not as a finished contract. Accepts file input via file_ref — no need to extract file content first.

Writing a schema by hand from a folder of sample payloads is slow, and it is easy to miss the one field that is sometimes a string and sometimes a number. You get a faithful first draft, and the disagreements are called out rather than smoothed over.

When to use it

Scope

Non-goals

• We never check a document against a schema — that is schema.validate_json
• We never invent constraints you cannot see in the examples, such as formats or ranges
• We never write conditionals, oneOf, or $ref; you add those yourself

Properties

Capabilities

Cost & billing

Flat rate per invocation

File handling

File limits

Constraints this skill applies to your input and generated files. The “Limited by” column shows the static source of each cap; none of these can be raised by upgrading your plan.

How we handle your data

Input schema

Output schema

API contract

curl -sS -X POST "$BASE/v1/execute" -H "Authorization: Bearer $KEY" -H 'Content-Type: application/json' -d '{"skillKey":"schema.infer_json_schema","input":{"examples":["{\"name\":\"Alice\",\"age\":30}","{\"name\":\"Bob\",\"age\":25}"]}}'

import { Squillify } from "@squillify/sdk";

const sq = new Squillify({ apiKey: process.env.SQUILLIFY_API_KEY });

const { output } = await sq.execute("schema.infer_json_schema", {"examples":["{\"name\":\"Alice\",\"age\":30,\"email\":\"[email protected]\"}","{\"name\":\"Bob\",\"age\":25}"]});

Errors

Other limits

Limitations

• We never check a document against a schema — that is schema.validate_json
• We never invent constraints you cannot see in the examples, such as formats or ranges
• We never write conditionals, oneOf, or $ref; you add those yourself
• Checking documents against a schema you already have (schema.validate_json)
• Writing schemas with conditional or polymorphic branches
• Comparing two schema versions (schema.diff_json_schema)

How to call

POST /v1/execute (requires authentication)

{
  "skillKey": "schema.infer_json_schema",
  "input": {
    "examples": [
      "{\"name\":\"Alice\",\"age\":30,\"email\":\"[email protected]\"}",
      "{\"name\":\"Bob\",\"age\":25}"
    ]
  }
}

Examples

{
  "examples": [
    "{\"name\":\"Alice\",\"age\":30,\"email\":\"[email protected]\"}",
    "{\"name\":\"Bob\",\"age\":25}"
  ]
}

{
  "schema": {
    "$schema": "https://json-schema.org/draft/2020-12/schema",
    "required": [
      "name",
      "age"
    ],
    "type": "object",
    "properties": {
      "name": {
        "type": "string"
      },
      "email": {
        "type": "string"
      },
      "age": {
        "type": "integer"
      }
    }
  },
  "warnings": [],
  "examplesAnalyzed": 2
}

{
  "examples": [
    "{\"value\":\"hello\"}",
    "{\"value\":42}"
  ]
}

{
  "schema": {
    "$schema": "https://json-schema.org/draft/2020-12/schema",
    "required": [
      "value"
    ],
    "type": "object",
    "properties": {
      "value": {
        "type": [
          "string",
          "integer"
        ]
      }
    }
  },
  "warnings": [
    "Mixed types detected: [string, integer]"
  ],
  "examplesAnalyzed": 2
}

Recommended use cases

• Start an API contract from payloads you already have
• Describe the shape of JSON coming out of a log or an export you did not design
• Spot fields whose type is inconsistent across real-world samples

Avoid for

• Checking documents against a schema you already have (schema.validate_json)
• Writing schemas with conditional or polymorphic branches
• Comparing two schema versions (schema.diff_json_schema)

Related skills

Composable with

Safety & privacy

• Runs without calling any external service.
• Your payloads are never stored permanently.
• Telemetry captures metadata only — never your payload contents.
• We never train models on your payloads.
• No human reviews your payloads.

Tags