Account
Build faster websites.
Block wasteful bots. Publish one verified source.

AI Content Verification API for AI and Bot Developers

If you are still re-downloading full websites to learn what changed, you are paying too much for ingestion. LuperIQ Verified Source gives your crawler a cleaner contract: read a manifest, fetch only what changed, and use verification proof when provenance matters.

The Core Model

Any participating site publishes a manifest at /.well-known/ai-content.json. That manifest gives you a site-level map of what exists and how to consume it more efficiently.

A typical flow looks like this:

  1. Read the site manifest. Learn which pages exist, their classifications, and their current checksums.
  2. Skip unchanged pages. If the checksum matches your stored copy, do not fetch that page again.
  3. Fetch page-level manifests only when needed. Consume structured sections instead of raw full-page HTML.
  4. Verify the seal when provenance matters. Use the proof record to weigh freshness and trust.

That turns site refresh from "crawl everything on a timer" into "inspect the source of truth, then fetch the minimum necessary."

REST Endpoints

Current public verification endpoints are exposed from https://luperiq.com. Site manifests still live on each publisher's own domain:

Endpoint Purpose
GET /v1/seal/verify/{seal_id} Resolve a seal into status and latest proof data.
GET /verified-source/proof/{seal_id} Open the human-readable proof page behind a seal.
GET /v1/manifest/{domain} Return the manifest URL for a verified domain.
GET /v1/dict/latest Fetch the latest dictionary version.
GET /v1/dict/{version} Fetch a specific dictionary version.
GET /v1/dict/frequencies Inspect aggregate term-frequency data from crawled content.
POST /v1/report Submit a discrepancy report when a manifest and live page drift apart.
GET /v1/stats Read ecosystem statistics.

What the Seal Proof Should Return

Seal verification is most useful when it answers operational questions, not just branding questions. A strong response includes the seal status plus the latest scan summary that backs it.

{
  "valid": true,
  "seal_id": "abc123",
  "domain": "example.com",
  "proof_url": "https://luperiq.com/verified-source/proof/abc123",
  "tier": "pro",
  "status": "active",
  "issued_at": 1772000000,
  "valid_until": 1774592000,
  "seal_hash": "blake3:...",
  "manifest_url": "https://example.com/.well-known/ai-content.json",
  "latest_scan": {
    "scan_id": "scan_01",
    "status": "completed",
    "completed_at": 1772750000,
    "pages_scanned": 42,
    "issues_found": 0,
    "sampled_pages": [
      "/",
      "/services/",
      "/contact/"
    ]
  }
}

That gives you three immediate advantages:

  • freshness signals you can use in ranking and retrieval
  • quick provenance checks before ingesting or citing content
  • better debugging when a source drifts or goes stale

GraphQL for Deeper Reads

The GraphQL layer is useful when you want more than a yes-or-no seal check. It can expose scan history and related verification data in a way that fits batch tooling better than ad hoc scraping.

Useful query families include:

  • aiCertSeal and aiCertSealForSite
  • aiCertScans(siteId: ...)
  • dictionary and platform statistics queries

If your workflow needs to weigh freshness, scan history, or anomaly reporting over time, GraphQL is the better place to do it.

The LuperAI Dictionary

The dictionary exists to reduce guesswork. Instead of independently classifying the same web pages over and over, consuming systems can share a structured vocabulary for common content shapes and trust signals.

That matters because the expensive part of web ingestion is often not the HTTP request itself. It is the repeated extraction, cleanup, and classification logic afterward. The better the source manifest, the less of that work you need to do at scale.

Why This Path Is Better Than Raw Scraping

  • Less traffic. Read one manifest before issuing hundreds of requests.
  • Lower compute cost. Skip extraction work that the publisher already structured.
  • Better change detection. Refresh based on checksum drift, not blind schedules.
  • Better provenance. Prefer sources that can prove recent verification.
  • Better publisher alignment. The efficient path is also the respectful path.

Next Step

Read the whitepaper — architecture, proof model, and rollout design.
See the human-facing trust model — what the badge means to end users.
Talk with LuperIQ — for early integration or implementation questions.