How an AI agent can call Ready Check to triage inbound email

If you build an agent that reads inbound mail, a general model can draft a reply. It is less reliable at the specialist work: comparing lookalike domains, reading attachment file names, and applying a fixed rule set. Ready Check is a prepaid API for that inspection step. It returns structured findings. It does not write a reply, connect to an inbox, or send mail.

This guide is for people who build AI agents. It covers what to send, how to call the live API, and what to do with the result. The full contract, prices, and check list are on the Ready Check facts page.

What Ready Check is for

You pass in the text of one email: sender, reply-to, subject, body, and attachment file names if you have them. Ready Check runs named rule checks first (the same families used in ReadyUpAgents email templates), then a model also inspects the email together with those check results.

You get back a kind, a verdict, scores, the named checks that fired, and short findings such as what the message looks like and what to watch for. Your own agent can use that data to draft in your employer’s style, with your review before anything is sent. Ready Check never connects to Gmail or any other inbox.

Optional kind values add specialist labels, for example sponsorship, form_lead, guest_post, key_request, podcast_guest, cold_pitch, pr_pitch, or security_report. If you leave kind out, the shared checks still run and you get a general verdict. Pictures, videos, audio, and file contents are not accepted; file names are signals only.

The call

The API is live at https://api.readyupagents.com. There is no account or dashboard. The agent generates its own prepaid key, buys credit once, and reuses that key. How to buy Ready Check credit and reuse a prepaid key covers key generation and checkout. A standard check costs $0.0025; an extra-long check costs two checks ($0.005).

Send a JSON POST:

POST https://api.readyupagents.com/v1/ready-check
Authorization: Bearer YOUR_READY_CHECK_KEY
Content-Type: application/json

{
  "kind": "sponsorship",
  "from": "Partnerships <deals@example.com>",
  "reply_to": "brandteam@example.com",
  "subject": "Paid collab, brief attached",
  "body": "The email text you already have from the inbox…",
  "attachment_names": ["Campaign_Brief.pdf"],
  "context": { "rate_card": { "youtube_integration": 1500 }, "currency": "USD" }
}

The body may be up to 16,000 characters. Serialized context is limited to 1,000 characters, attachment names to 30, and the raw request to 64 KiB. Longer emails may count as two checks. Oversized input is rejected before the model runs, at no charge. An empty body is rejected with HTTP 400. A key with no credit, or with too little credit for a two-check charge, returns 402 and is not charged.

Send an optional Idempotency-Key header (1–128 characters of A-Za-z0-9._-) so a retry after a network timeout is not double-charged. Without that header, do not blindly retry after a timeout; keep the request_id from a successful or pending response if you have one.

What to do with the result

Treat the result as a sorting aid, not a guarantee. A careful scammer can pass checks, and a real sender can trip one. Read verdict, checks_fired, watch_for, and any specialist fields before your agent writes. Ready Check does not include a draft body. If a rule already raised a warning, do not let your own model talk itself out of that warning.

Typical agent flow:

Your workflow already has the email text (from the user’s own inbox connection, not from Ready Check).
Call Ready Check and wait for the findings.
Your agent may draft a reply from those findings, in the user’s voice, and leave it for review. Nothing is sent by Ready Check.

For the response fields, kinds, and what “findings only” means in practice, see What Ready Check returns: findings, not drafts. For prices, prepaid keys, refunds, and the full check list, use the facts page. An agent-readable version of this guide is at Ready Check for agents: how to call it.

ReadyUpAgents is a verified n8n creator: https://n8n.io/creators/readyupagents/.

ReadyUpAgents is independent and not affiliated with or endorsed by n8n, Google, OpenAI, Stripe or Cloudflare. Product and company names belong to their owners.