VERSION 1 · REST + JSON · INCLUDED IN PRO / TEAM / AGENCY · STANDALONE API + REPORTS
FOR AI VISIBILITY TOOLS · PUBLIC DOCUMENTATION
The technical layer alongside visibility tracking.
For teams evaluating integrations at tools such as Otterly and Peec: AI-Primed exposes the website checks, observed evidence and generated fix drafts your product can read through a scoped API key. These are prospective integration examples—not partners, customers or existing integrations.
Submit a customer’s public site, retrieve its readiness report, and bring the saved robots.txt, llms.txt or JSON-LD drafts into a review workflow where available. The audit checks reachability and readability; it does not prove why a brand was omitted from an AI answer or promise future mentions. Public documentation does not mean unauthenticated access: each key can read only its own account’s reports.
Submit a public domain, retrieve six-category scores and observed findings, then retrieve saved fix files. Keep the technical evidence alongside the score—not behind another dashboard.
POST/auditsQueue an audit
GET/audits/{id}Retrieve scores and evidence
GET/audits/{id}/fixesRetrieve generated files
GET/domains/{domain}/scoreRead a saved observation
One export contract
The API and JSON report download share agentready.audit/v1. Scores, findings, fixes and fetch evidence travel together.
Keys with boundaries
Named read-only or audit-creation keys. Hashed at rest, shown once, rotatable and revocable. Reads stay within the key’s account.
Bounded by design
Shared account quotas, daily submission caps and concurrency limits. Saved score lookups never start a new crawl. Public-page fetching uses the audit’s safety checks.
OUR OWN SITE · RECORDED TEST OBSERVATION
We put our own site through the API.
ai-primed.lovable.app returned a completed audit with two low-severity findings and an Organization schema draft. One missing identity signal remained; we did not invent public profiles to fill it.
A keyed API audit was also compared with the dashboard JSON export: scores, findings, fixes and evidence matched exactly. Processing was invoked locally; this is not deployed scheduling proof, a customer endorsement, or evidence of AI mentions.
SIGNED EVENTS · ACTIVATION PENDING
Audit completed. Score changed. Your workflow notified.
The implemented webhook service signs audit.completed and score.changed events with HMAC-SHA256. An isolated sender validates public addresses, pins the connection and checks the receiver’s certificate before sending. No redirects, no internal-network destinations, and up to five attempts with delivery history.
Eligible API accounts can opt in after protected delivery is activated. Score-change events require two scored observations; temporary failures are not score changes. Existing destinations are never opted in automatically. Deployment and a real receiver test remain pending—poll for results in the meantime.
API access is included in every paid platform plan — Pro, Team and Agency — sized to that plan's sites. Developers who only need the API can buy API + Reports, shown under the plans above. Free has no API access.
The allowance that matters is site audits per month: each new audit you submit counts once. Reading stored results is a separate, larger allowance and never starts a crawl. Generated remediation is a draft to review and apply — not automatic site changes or a guaranteed score improvement.
Scheduled monitoring is metered separately and never draws on your API allowance. Even so, every tier's submission allowance exceeds what daily monitoring of all its sites would use, leaving the rest for re-tests, competitor checks and onboarding:
Pro: 5 sites × 30 days = 150 monitoring · 500 submissions included · 350 left for re-tests, competitor checks and onboarding
Team: 25 sites × 30 days = 750 monitoring · 2,000 submissions included · 1,250 left for re-tests, competitor checks and onboarding
Agency: 50 sites × 30 days = 1,500 monitoring · 5,000 submissions included · 3,500 left for re-tests, competitor checks and onboarding
All keys share account allowances. Minute limits apply per key and account; submission and concurrency caps apply per account. Accepted submissions count even if fetching fails. Existing site limits also apply. Preview purchases use test payments; live payment verification remains outstanding.
What is—and isn’t—verified
Test checkout, account access, audit submission, local processing and result retrieval were verified. Deployed scheduled processing and live checkout remain unverified. Signed audit-completion and score-change delivery is built with public-IP connection pinning, certificate verification and bounded retries. The protected sender still needs deployment and a real receiver test; webhook activation is pending, so use polling for now. Generated files are drafts, not applied or verified fixes. AI-Primed does not control assistant rankings or mentions.