Policy evidence before web ingestion

Know why a URL entered your AI data pipeline

AgentPermission checks publisher policy signals before your crawler, RAG loader, extraction job, or training pipeline fetches a URL. Get allow, deny, or escalate with the source evidence behind the decision.

Create Free Ingestion KeyRun a Live Policy Check
npx -y agent-permission@latest install --adapter codexAdd before fetch
Quick Tests:
Integration Snippet
curl -X POST https://agent-permission-api-788656703262.us-central1.run.app/v1/check \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer apk_test_key" \
  -d '{
    "url": "https://example.com",
    "action": "summarize"
  }'

How it works

One checkpoint before content becomes data

Put the check immediately before a crawler or data loader fetches and stores an external URL.

  1. 01

    Name the action

    Classify the ingestion purpose as crawl, summarize, extract, or train.

  2. 02

    Read the available signals

    Read robots.txt, explicit Content Signals, OpenTerms, and any explicit permission directives in llms.txt.

  3. 03

    Attach the evidence

    Continue on allow, stop on deny, review escalate, and store the receipt ID with the ingestion job.

What it gives you

Evidence your pipeline can retain

  • A decision tied to an exact URL and ingestion purpose
  • The reasons and source documents used
  • A service-signed receipt with URL and source hashes

What it does not claim

Legal or account authorization

  • It is not legal advice or proof of legal compliance
  • It does not prove that a publisher owns every implicated right
  • It does not authorize posting, purchasing, or account changes
Read the exact trust boundary →

Pricing

Evaluate on one pipeline, then move to production

Every plan uses the same conservative decision engine and evidence format. Start free without a credit card.

Evaluation

Validate the ingestion checkpoint with 1 active API key and 100 checks/month.

$0/mo
  • 100 checks / month
  • 1 Active API Key
  • CLI, UI, and API checks
  • Source evidence and signed receipts
Create Free Ingestion Key
POPULAR

Production

For web-ingestion pipelines that need production volume, multiple keys, and evidence attached to every decision.

$29/mo
  • 50,000 checks / month
  • 3 Active API Keys
  • CLI, UI, and API checks
  • Source evidence, receipts, and usage dashboard
Get Started

Team

For data teams operating multiple higher-volume ingestion workflows.

$99/mo
  • 250,000 checks / month
  • 10 Active API Keys
  • CLI, UI, and API checks
  • Source evidence, receipts, and usage dashboard
Upgrade to Team

Production design partners

Running third-party web ingestion in production?

We are working directly with a small number of crawler, RAG, and data-governance teams to validate retention, evidence export, policy-change monitoring, and procurement requirements before promising them as product features.

Request a Pipeline ReviewReview the Trust Boundary

Receipt boundary

A record of what was checked—not a legal verdict

Receipts use a service-held HMAC signing key and can be validated through the public verification endpoint or CLI.

Decision
{
  "id": "chk_235344b4",
  "decision": "allow",
  "confidence": "high",
  "risk": "low",
  "reasons": [
    "OpenTerms explicitly allows read_content"
  ]
}
Service-signed receipt
{
  "receipt_id": "rcpt_a842ae72",
  "check_id": "chk_235344b4",
  "url_hash": "sha256…",
  "action": "summarize",
  "decision": "allow",
  "source_hashes": ["sha256…"],
  "issued_at": "2026-07-18T12:00:00.000Z",
  "account_id": "acct_…",
  "signature_alg": "HMAC-SHA256",
  "signature": "service-signature"
}

Because the signing key stays with AgentPermission, validation is service-backed rather than independently offline. See the API contract →