ARD ai-catalog.json Validator

What does ARD Catalog Validator do?

Agent selection and quickstart

Use when: Validate an ARD ai-catalog.json before integration or return diagnostics for repair automation.

Input: Exactly one inline catalog object or public HTTPS url.

Returns: valid, errorCount, errors[].code, and errors[].path.

Limits: Returns diagnostics without applying repairs. ARD draft 0.9/catalog 1.0 is pinned.

Cost: $0.01 per completed determination, including a completed negative result; operational failures are not billed. See the live Pricing tab for current rates.

{
  "catalog": {
    "specVersion": "1.0",
    "entries": [
      {
        "identifier": "weather",
        "displayName": "Weather",
        "type": "application/json",
        "url": "https://example.com/weather"
      }
    ]
  },
  "validationMode": "strict"
}

Verified output excerpt for this input (selected fields; canary XoqDLR5oIZVa2hvNt):

{
  "contractVersion": "1.0",
  "ok": true,
  "valid": false,
  "errorCount": 1,
  "errors": [
    {
      "code": "INVALID_IDENTIFIER",
      "path": "/entries/0/identifier",
      "severity": "error",
      "message": "The ARD resource identifier is invalid.",
      "expected": "An identifier matching urn:air:<publisher>:<namespace>:<name>",
      "actual": "weather",
      "suggestion": "Use a valid ARD identifier beginning with urn:air: and containing the required components."
    }
  ]
}

Open this public example Task. Live sources can change; this records the checked example rather than guaranteeing a future result.

The result is in the default dataset. With Apify MCP, inspect this Actor using fetch-actor-details, invoke call-actor, then read the returned dataset with get-dataset-items. Run limits belong in callOptions, not in Actor input.

Agent search queries

• "validate ai-catalog.json"
• "ARD catalog validation errors"
• "check Agentic Resource Discovery conformance"
• "machine readable ARD repair paths"
• "validate an inline ARD manifest"

ARD Catalog Validator determines whether an Agentic Resource Discovery catalog conforms and returns exactly what a coding agent must fix. Supply either a raw JSON object for fully offline validation or a public HTTPS URL. The Actor returns stable rule codes, JSON Pointer paths, expected and actual values, severities, and deterministic repair suggestions through Apify API, schedules, integrations, or hosted MCP tooling.

It uses the same pinned schema and semantic engine as ARD Catalog Resolver, with no LLM, browser, paid API, or automatic mutation of your document.

Why use ARD Catalog Validator?

The official ARD project provides a free conformance CLI. This paid primitive is for workflows that need an independently deployable Apify/MCP contract, one versioned JSON result, stable diagnostics rather than Ajv-specific prose, remote URL safety, and parity with the Resolver’s production interpretation. Typical agent queries include “Is this ARD catalog valid?”, “Why does this ai-catalog.json fail?”, and “Return machine-readable repairs for this ARD manifest.”

How to use ARD Catalog Validator

1. Open the Input tab.
2. Paste a catalog JSON object or provide one public HTTPS url, never both.
3. Select strict mode for conformance or permissive mode while investigating extensions.
4. Run and consume the single dataset envelope.

Input

{
    "catalog": {
        "specVersion": "1.0",
        "entries": [{ "identifier": "weather", "displayName": "Weather", "type": "application/json", "url": "https://example.com/weather" }]
    },
    "validationMode": "strict"
}

Raw input causes no network request. Strict mode rejects fields that the pinned upstream schema forbids. Permissive mode demotes only those additionalProperties failures to warnings; entry extension fields that the ARD schema explicitly permits remain valid in both modes. An unsupported catalog version remains invalid.

Output

{
  "contractVersion": "1.0",
  "ok": true,
  "valid": false,
  "errorCount": 1,
  "errors": [
    {
      "code": "INVALID_IDENTIFIER",
      "path": "/entries/0/identifier",
      "severity": "error",
      "message": "The ARD resource identifier is invalid.",
      "expected": "An identifier matching urn:air:<publisher>:<namespace>:<name>",
      "actual": "weather",
      "suggestion": "Use a valid ARD identifier beginning with urn:air: and containing the required components."
    }
  ]
}

You can download the dataset in formats such as JSON, HTML, CSV, or Excel; JSON is recommended for repair automation.

Data table

Field	Meaning
valid	Conformance in the selected mode.
ardSpecVersion	Pinned ARD prose draft version.
catalogSpecVersion	Manifest version declared by the input.
schemaRevision	Exact upstream Git revision.
errors	Repair-grade blocking issues.
validationWarnings	Non-blocking issues.

Pricing / Cost estimation

The published price is $0.01 for one catalog-validation event. Valid and invalid completed determinations are billable. Input rejection and URL network/resource/internal failures are free. Check the live Actor details for current pricing.

Tips and limits

Prefer inline input in CI and code-generation workflows: it is faster, deterministic, and removes network uncertainty. The Actor never emits a patched catalog, so downstream automation must apply and review its own changes. Remote URL responses are limited to public HTTPS, bounded redirects, and 1 MiB decoded content. DNS is resolved and pinned per request, private/local/link-local targets are blocked again after redirects, and URL query values are redacted from result provenance and logs.

Invalid JSON and schema-invalid catalogs are completed validation determinations (ok: true, valid: false). Invalid Actor input and URL retrieval failures return ok: false with a stable machine error and are not charged.

FAQ, disclaimers, and support

The supported baseline is ARD draft 0.9 and catalog manifest 1.0 at schema revision 5fa2f5aef790b478319f6a3b43adf4661b0ed0e0. This is not a replacement for reviewing the official specification or free conformance tooling. Use the Actor’s Issues tab for a diagnostic that differs from the pinned upstream behavior, including a minimal anonymized catalog.

Related Actors and next steps

• Find and normalize a domain catalog before validating it

Public examples by task

• Validate an ARD catalog before integration — Validate a live ARD catalog and get structured conformance errors and repair guidance, using Ora.ai as a valid example.
• Get repair paths for an invalid ARD catalog — Validate a deliberately invalid ai-catalog.json and get the stable INVALID_IDENTIFIER diagnostic, JSON Pointer path, and repair suggestion.