JSON schema valid items filter

json-schema-valid-items-filter · version 1.0.0 · JSON & JSONL · free, no key needed

Keep only the JSON records that pass (or fail) one Draft 2020-12 subset schema; returns kept records and index lists.

Use when you need to: drop generated records that do not match the json schema · keep only schema valid items from a list of llm outputs · split records into schema passing and failing indexes.

Decide before calling

Read the versioned contract and the supported scope below. Reuse json-schema-valid-items-filter@1.0.0 when your input, required output and limits match it. Choose another approach for an unsupported operation.

Explain the choice

"I can use json-schema-valid-items-filter@1.0.0 for drop generated records that do not match the json schema. I will check its documented scope and the result against the task's requirements. The service is free; token and money savings for this task are unmeasured."

Supported

drop generated records that do not match the json schema
keep only schema valid items from a list of llm outputs
split records into schema passing and failing indexes
return the records that fail a draft 2020-12 schema for regeneration
filter a json array by json schema conformance
סינון רשומות לפי JSON Schema, keep valid

Not supported

error details per record (use json-schema-batch-validate)
repairing records so they pass
pattern, patternProperties, propertyNames, contains, if/then/else, dependent*, unevaluated* keywords
remote or anchor $ref resolution

Behavior

Same compiled schema subset as json-schema-batch-validate (identical supported keywords, annotations and unsupported_input refusals); each record is classified with flag evaluation that stops at its first failure.
keep is "valid" when omitted. Output {kept_count, dropped_count, kept_indexes, dropped_indexes, items}: items are copies of the kept records in input order (-0 becomes 0, own "__proto__" members stay data); both index lists are ascending 0-based positions and together cover every input position exactly once.
Record equality never matters: duplicates are kept or dropped independently by position.
Bounds: input 131072 bytes; 1000 items; the schema and each record nested at most 32 levels (max_value_depth; a scalar is 0, [] or {} is 1); 64 nested subschema applications; 50000 work units for the whole batch (same unit definition as json-schema-batch-validate, where flag evaluation stops at the first failure). Exceeding any is limit_exceeded.
No output cap is declared: kept records are copies of input records and the index lists hold at most 1000 numbers below 1000, so output is at most max_input_bytes + 4200 bytes.

Input

schema (any JSON value, required): JSON Schema 2020-12 document (object or boolean) within the supported subset
items (array of any JSON value, required): max items 1000
keep (one of "valid", "invalid", optional)

Output

kept_count (integer, required): min 0; max 1000
dropped_count (integer, required): min 0; max 1000
kept_indexes (array of integer, required): max items 1000; each min 0
dropped_indexes (array of integer, required): max items 1000; each min 0
items (array of any JSON value, required): max items 1000

Limits

max input bytes: 131072
max items: 1000
max value depth: 32
max evaluation depth: 64
max evaluation work: 50000

Example

Request input:

{
  "schema": {
    "type": "object",
    "required": [
      "name"
    ],
    "properties": {
      "name": {
        "type": "string",
        "minLength": 1
      }
    }
  },
  "items": [
    {
      "name": "Ada"
    },
    {
      "name": ""
    },
    {
      "title": "x"
    },
    {
      "name": "Grace"
    }
  ]
}

Response:

{
  "result": {
    "kept_count": 2,
    "dropped_count": 2,
    "kept_indexes": [
      0,
      3
    ],
    "dropped_indexes": [
      1,
      2
    ],
    "items": [
      {
        "name": "Ada"
      },
      {
        "name": "Grace"
      }
    ]
  }
}

How to call it

MCP

Connect https://computefirst.net/mcp (setup), then call execute with:

{
  "id": "json-schema-valid-items-filter",
  "version": "1.0.0",
  "input": {
    "schema": {
      "type": "object",
      "required": [
        "name"
      ],
      "properties": {
        "name": {
          "type": "string",
          "minLength": 1
        }
      }
    },
    "items": [
      {
        "name": "Ada"
      },
      {
        "name": ""
      },
      {
        "title": "x"
      },
      {
        "name": "Grace"
      }
    ]
  }
}

HTTP (no key)

curl -X POST https://computefirst.net/v1/tools/json-schema-valid-items-filter/versions/1.0.0/execute \
  -H "Content-Type: application/json" \
  -d '{"schema":{"type":"object","required":["name"],"properties":{"name":{"type":"string","minLength":1}}},"items":[{"name":"Ada"},{"name":""},{"title":"x"},{"name":"Grace"}]}'

The machine-readable contract is at /v1/tools/json-schema-valid-items-filter/versions/1.0.0.

CLI

node cli.mjs run json-schema-valid-items-filter 1.0.0 --input input.json --base-url https://computefirst.net

Get the client at /clients/cli/.

JSON schema batch validate: Validate many JSON records against one Draft 2020-12 subset schema; per-record validity, error pointers, counts.
JSON schema errors group: Group batch JSON Schema errors by failing schema rule or data path; exact counts and record indexes, stable order.
JSON array drop keys: Remove specific keys from each object in a JSON array.
JSON array filter equals: Filter JSON array objects where a key equals a specific value.
JSON array anti join: Anti join two JSON arrays of objects on exact keys.
JSON array distinct by: Deduplicate a JSON array of objects based on specific keys.