JSONL Schema Validator

updated 12 May 2026

JSONL Schema validator. Check every record in a JSONL (JSON Lines) file against a JSON Schema you supply. Reports each failing line with its JSON pointer and the rule it broke. Up to 1 GB, runs in your browser.

Before you start

You need two things: a JSON Schema (Draft-07 or later, paste it into the schema pane) and a JSONL data file. The validator runs every record through the schema and reports each violation with its JSON pointer.

How to use it

1. Paste your JSON Schema into the left pane.
2. Paste or drop your JSONL data on the right.
3. Click Validate.
4. Read the error list — each entry says which line failed, which JSON pointer (e.g. /email or /items[2]/sku), and the rule it broke.
5. Click Download passing rows if you want to keep only the records that satisfy the schema.

Supported keywords

• type: object, array, string, integer, number, boolean, null; supports union types as arrays.
• required: list of required keys on an object.
• properties + additionalProperties: false for strict objects.
• items for array element validation.
• enum, const for fixed-value checks.
• pattern (regex) on strings.
• minLength / maxLength on strings.
• minimum / maximum on numbers.
• minItems / maxItems on arrays.
• format: "email" for a quick email check.

Complex features like $ref, oneOf, anyOf compositions aren't supported here — use a full validator like Ajv for those.

Example

Schema:

{
  "type": "object",
  "required": ["id", "email"],
  "properties": {
    "id":    { "type": "integer" },
    "email": { "type": "string", "format": "email" },
    "age":   { "type": "integer", "minimum": 0 }
  },
  "additionalProperties": false
}

Data:

{"id": 1, "email": "[email protected]", "age": 33}
{"id": 2, "email": "not-an-email"}
{"id": 3, "email": "[email protected]", "extra": true}

Reports: line 2 — /email not a valid email; line 3 — /extra additional property not allowed.

Tips & common pitfalls

• Need a schema? Run your file through the Schema inferrer first — it generates a starter Draft-07 schema you can tighten up.
• "additionalProperties: false" is strict. Without it, extra keys are tolerated. Enable it only if your downstream system is strict too.
• Use this in CI conceptually. A failing schema check before you train an AI model can save hours of wasted compute.

Frequently asked questions

Related tools

• Schema inferrerGenerate a schema from your data first.
• JSONL ValidatorPure syntax check, no schema needed.
• Auto-fixerRepair broken JSONL before validating.
• OpenAI fine-tune validatorProvider-specific schema checks for chat-format files.