Run AEO checks in your pipeline

Open Settings → API keys, click New API key, give it a label like ci-deploy, and copy the aeol_… token shown once. Add it as a secret in your CI provider (GitHub Actions: AEOLENS_API_KEY; GitLab: a masked variable).

API access ships on Growth ($99/mo) and Agency ($249/mo) plans.

The endpoint runs synchronously: it queues the scan, waits up to max_wait_seconds, and returns a pass/fail body. Run it against the URL you just deployed (a staging URL is the common pattern).

curl -X POST https://aeolens.ai/api/v1/ci/scan \
  -H "Authorization: Bearer $AEOLENS_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://staging.acme.com",
    "min_score": 70,
    "max_wait_seconds": 180
  }'

Response (status 200):

{
  "status": "complete",
  "passed": true,
  "score": 78,
  "threshold": 70,
  "scan_id": "0a1b2c3d-…",
  "scan_url": "https://aeolens.ai/scan/0a1b2c3d-…",
  "failed_checks": ["SCHEMA", "OG"],
  "error": null,
  "wait_seconds": 12.4
}

The endpoint always returns 200 on a successful scan — even when the threshold isn't met. Your script reads .passedand decides whether to fail the build. This keeps "AEOlens errored" distinguishable from "deploy was blocked."

Drop this into .github/workflows/aeo.yml (or merge into your existing deploy workflow). It runs after deploy and fails the job if AEOlens reports passed: false.

name: AEO gate

on:
  deployment_status:
    states: [success]

jobs:
  aeo-scan:
    runs-on: ubuntu-latest
    steps:
      - name: Run AEOlens scan
        env:
          AEOLENS_API_KEY: ${{ secrets.AEOLENS_API_KEY }}
          DEPLOY_URL: ${{ github.event.deployment_status.target_url }}
        run: |
          RESPONSE=$(curl -sf -X POST https://aeolens.ai/api/v1/ci/scan \
            -H "Authorization: Bearer $AEOLENS_API_KEY" \
            -H "Content-Type: application/json" \
            -d "{\"url\":\"$DEPLOY_URL\",\"min_score\":70}")
          echo "$RESPONSE" | jq .
          PASSED=$(echo "$RESPONSE" | jq -r .passed)
          if [ "$PASSED" != "true" ]; then
            echo "::error::AEO gate failed — see scan_url in the response."
            exit 1
          fi

Drop into .gitlab-ci.yml. Set AEOLENS_API_KEY and DEPLOY_URL as project variables.

aeo_gate:
  stage: post-deploy
  image: curlimages/curl:latest
  script:
    - |
      RESPONSE=$(curl -sf -X POST https://aeolens.ai/api/v1/ci/scan \
        -H "Authorization: Bearer $AEOLENS_API_KEY" \
        -H "Content-Type: application/json" \
        -d "{\"url\":\"$DEPLOY_URL\",\"min_score\":70}")
      echo "$RESPONSE"
      echo "$RESPONSE" | grep -q '"passed":true' || exit 1
  rules:
    - if: $CI_COMMIT_BRANCH == "main"

The CI endpoint consumes a scan from your plan's monthly quota — it isn't a separate bucket. Growth covers unlimited scans with a fair-use cap of 15/day per workspace; Agency covers 50/day. If you scan on every PR you'll want Growth or above.

Hitting the cap returns a 402 with code: "QUOTA_EXCEEDED" rather than the 200 pass/fail shape.