ALPNAI / FIELD NOTES / 14.09.2026

Free evidence structure checker for AI agents

Inspect sources, dates and documented fields in a JSON object. This check evaluates the structure of evidence, never the truth of its claims.

What does this tool check?

Valid JSON can still be unusable for a research agent. This checker is designed to identify whether a source, a date and populated or explicitly unknown fields are present. Its result describes the submitted structure. It does not automatically read the cited publications and does not certify that an event, price or date matches reality. Structural completeness is only one step.

What should you prepare?

Use a public example of your response containing a claim, a source link and the dates available to you. Keep an unestablished field null and explain the scope of that uncertainty. A delivery timestamp should not replace a publication date. Leave out secrets, access keys and confidential information: none of them is necessary to examine whether the format contains useful evidence fields.

How should you read the result?

A supplied source can be fabricated or unrelated to the claim. A correctly formatted date can be old. A populated field can contain an error. The check therefore reports present or missing elements rather than a financial reliability score. After inspecting the structure, open the original source and compare its content with each claim your agent intends to use.

Use it as an integration starting point

This free utility can help prepare a response for a workflow, compare evidence formats and document unknown fields. Examining a structure does not require buying a catalog product. Availability follows the integration of the tool; it enables neither cryptocurrency settlement nor continuous monitoring. For the next step, read the pilot request contract, inspect its sample and check its actual coverage before requesting access.

Check a sample structure Activate my agent

Check an evidence structure

Paste a JSON document. Analysis stays in your browser and does not verify the truth of claims.

JSON

{
  "schema_version": "1.0",
  "snapshot_id": "openai-2026-09-14-v1",
  "entity": "OpenAI",
  "compiled_at": "2026-09-14",
  "source_event_date": "2026-06-08",
  "coverage": "Curated primary-source collection. Not exhaustive; not real-time.",
  "status": "confidential_draft_s1_submission_announced",
  "status_scope": "The June 8, 2026 announcement confirms confidential submission of a draft registration statement.",
  "ipo_date": null,
  "ticker": null,
  "offer_price": null,
  "unknown_scope": "No timetable was decided in the cited announcement. These null fields do not assert the absence of subsequent announcements.",
  "facts": [
    {
      "id": "s1-submitted",
      "statement": "OpenAI announced the confidential submission of a draft S-1 to the SEC.",
      "source_id": "openai-s1-2026",
      "event_date": "2026-06-08",
      "type": "issuer_statement"
    },
    {
      "id": "timing-undecided",
      "statement": "The announcement said that the timing had not been decided.",
      "source_id": "openai-s1-2026",
      "event_date": "2026-06-08",
      "type": "issuer_statement"
    }
  ],
  "sources": [
    {
      "id": "openai-s1-2026",
      "title": "Confidential submission of draft S-1 to the SEC",
      "url": "https://openai.com/index/openai-submits-confidential-s-1/",
      "publisher": "OpenAI",
      "published_at": "2026-06-08"
    },
    {
      "id": "openai-structure",
      "title": "Our structure",
      "url": "https://openai.com/our-structure/",
      "publisher": "OpenAI",
      "published_at": null
    }
  ],
  "affiliation": "Independent service. Not affiliated with OpenAI.",
  "financial_advice": false
}

Sources

OpenAI — Dated primary-source example

SEC — Developer Resources and source access

Continue reading

OpenAI IPO evidence: what can an agent actually establish?

Connect an MCP agent with budgets and replayable purchases