Audit every client for AI citation.
Ship your own report.

# Audit a whole book of client domains for AI-citation readiness.
# One loop, one bearer token, structured findings you can roll into a report.
export OLLAGRAPH_API_KEY="osk_..."

for DOMAIN in clienta.com clientb.com clientc.com; do
  echo "== $DOMAIN =="

  # 1. The headline number: how cite-ready is the page for answer engines?
  curl -s -X POST https://api.ollagraph.com/v1/aeo/citation-readiness \
    -H "Authorization: Bearer $OLLAGRAPH_API_KEY" \
    -H "Content-Type: application/json" \
    -d "{\"url\": \"https://$DOMAIN\"}"

  # 2. Can the AI crawlers even reach it? GPTBot / ClaudeBot / PerplexityBot.
  curl -s -X POST https://api.ollagraph.com/v1/aeo/ai-bot-allowlist \
    -H "Authorization: Bearer $OLLAGRAPH_API_KEY" \
    -d "{\"url\": \"https://$DOMAIN\"}"

  # 3. Is the structured data answer engines look for actually present?
  curl -s -X POST https://api.ollagraph.com/v1/aeo/schema-coverage \
    -H "Authorization: Bearer $OLLAGRAPH_API_KEY" \
    -d "{\"url\": \"https://$DOMAIN\"}"
done

# White-label pipeline: call the audit endpoints, render with YOUR branding.
# The client sees your agency's report — they never need to know Ollagraph exists.
import os, httpx

API = "https://api.ollagraph.com"
client = httpx.Client(
    headers={"Authorization": f"Bearer {os.environ['OLLAGRAPH_API_KEY']}"},
    timeout=60.0,
)

def audit_client(domain: str) -> dict:
    url = f"https://{domain}"
    return {
        # One call, the full scored page audit (multiple AEO checks at once).
        "page_audit": client.post(f"{API}/v1/aeo/page-audit",
            json={"url": url}).json(),
        # Does the page lead with the answer? Heading hierarchy + snippet shape.
        "citation": client.post(f"{API}/v1/aeo/citation-readiness",
            json={"url": url}).json(),
        # Schema.org coverage gap analysis — what's missing for rich answers.
        "schema": client.post(f"{API}/v1/aeo/schema-coverage",
            json={"url": url}).json(),
        # Classic technical-SEO meta audit for the same report.
        "meta": client.post(f"{API}/v1/seo/meta-audit",
            json={"url": url}).json(),
    }

# Roll each result into your own PDF / Notion / Looker template.
# JSON in, your-brand report out. One workspace per client keeps it tidy.
report = audit_client("client.example.com")

# "Will ChatGPT actually see this page the way the client thinks?"
# The LLM fetch simulator answers that. Pair it with the freshness signal
# so the deliverable shows both reachability AND recency.
export OLLAGRAPH_API_KEY="osk_..."

# Simulate how an answer-engine crawler fetches and parses the page.
curl -X POST https://api.ollagraph.com/v1/aeo/llm-fetch-simulator \
  -H "Authorization: Bearer $OLLAGRAPH_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"url": "https://client.example.com/pricing"}'

# Audit the client's llms.txt — the file answer engines increasingly read.
curl -X POST https://api.ollagraph.com/v1/aeo/llms-txt-audit \
  -H "Authorization: Bearer $OLLAGRAPH_API_KEY" \
  -d '{"url": "https://client.example.com"}'

# Freshness: are the signals an answer engine uses to judge recency present?
curl -X POST https://api.ollagraph.com/v1/aeo/freshness-signal \
  -H "Authorization: Bearer $OLLAGRAPH_API_KEY" \
  -d '{"url": "https://client.example.com/blog/guide"}'