On this page
Getting started

Introduction

One key, one wallet, USD per report, no subscriptions.

One call crawls a site, reads its robots.txt and sitemap, compares it with up to two competitors' public sitemaps, and returns a ranked fix list with evidence URLs. Pay per call from a prepaid USD wallet; every response states cost_usd.

Quickstart

BASE=https://audit.45-94-66-237.sslip.io
KEY=$(curl -s -X POST $BASE/v1/keys -H "Content-Type: application/json" -d '{"label":"me"}' | jq -r .key)

# free: validates input and returns the price
curl -s -X POST $BASE/v1/quote/seo.audit -H "Content-Type: application/json" -d '{"domain":"yoursite.com"}'

curl -s -X POST "$BASE/v1/run/seo.audit?fields=summary,actions" \
  -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"domain":"yoursite.com","competitors":["rival.com"]}' | jq '.cost_usd, .output.summary'

Send the key as Authorization: Bearer <key> or X-API-Key: <key>. Keys are stored hashed and shown once.

Data modes

ModeSet withWhat is real
site (default, free)ANYAPI_PROVIDER=siteEverything in the report: fetched from the audited site, its robots.txt and sitemap, competitors' public sitemaps/homepages, and optionally Google PageSpeed Insights. No keyword volumes or backlinks; those sections are omitted and listed under not_covered.
mockANYAPI_PROVIDER=mockFictional keyword/backlink data, labelled demo: true. For UI testing only.
dataforseoANYAPI_PROVIDER=dataforseo + credentialsAdds licensed keyword/backlink data on top of the site crawl. Experimental; needs written resale/derivative approval before selling.

Run envelope

{"ok": true, "request_id": "req_…", "endpoint": "seo.audit", "status": "succeeded",
 "cost_usd": 1.0, "items": 1, "latency_ms": 2140,
 "source": {"id": "site", "name": "First-party crawl", "kind": "first_party", "demo": false},
 "output": { … the report … }}

Query controls (no change in cost): fields, max_items, summary=true; max_cost_usd refuses a call that would cost more. Prefer: respond-async or Prefer: wait=N returns 202 + Location to poll. See running and durable requests.

Site audit seo.audit $1.00

POST /v1/run/seo.audit
{
  "domain": "yoursite.com",          // required; URL or bare domain
  "competitors": ["rival.com"],      // optional, max 2
  "location": "us", "language": "en", "focus": "all"
}

What the free crawl checks

  • Crawlability & indexing: HTTPS, http→https and www/non-www redirects, redirect chains, robots.txt (RFC 9309 semantics), Googlebot blocked, sitemap present/declared/valid, broken pages and internal links, sitemap URLs that error, noindex (meta and X-Robots-Tag), missing or cross-domain canonicals.
  • On-page: missing/duplicate/long titles, missing/duplicate meta descriptions, missing/multiple H1, viewport, lang, image alt text, mixed content.
  • Content: thin pages (<250 words), JSON-LD structured data, Open Graph, competitor section and topic gaps from public sitemaps, content inventory vs competitors.
  • Performance: server response time, compression, HSTS; with ANYAPI_PAGESPEED=true, mobile Lighthouse score, LCP and CrUX field data from PageSpeed Insights.

Bounds: up to ANYAPI_CRAWL_MAX_PAGES pages (default 15, sampled across site sections), 4 parallel requests, 60 s budget, 2 MB per page. Only public IP addresses are fetched.

Response (main fields)

FieldMeaning
summaryScore 0–100, diagnosis, top opportunities/risks, site_scores (crawl_index, on_page, content, performance).
actions[]Ranked plan: priority, action (what to do), title (finding), why, impact, effort, affected, evidence (URLs).
planPriorities grouped into now, next, later.
technical_issues[]Every finding with severity, category, count, affected URLs and fix.
content_gaps[]Sections/topics competitors publish that this site does not (candidate topics; validate demand).
siteCrawl evidence: pages fetched, robots, sitemap, variants, homepage, PageSpeed, competitor snapshots.
not_covered, sources, meta.errors, partialWhat the report cannot tell you, where every number came from, and what failed.

Topic / keyword gap seo.keyword_gap $0.25

POST /v1/run/seo.keyword_gap
{"domain":"yoursite.com","competitor":"rival.com","limit":25}

In site mode this compares the two sites' public sitemaps (method: "sitemap_topic_gap"): sections and page topics the competitor has that you don't. With a licensed provider it compares ranked keywords with volume and difficulty.

Content brief seo.content_brief $0.50

POST /v1/run/seo.content_brief
{"keyword":"emergency plumber leeds"}

Needs keyword and SERP data. In free site mode it returns 501 not_available and does not charge.

Platform calls

CallPurpose
GET /v1/catalog, /v1/catalog/search?q=, /v1/catalog/{id}Endpoints, schemas, prices, source, measured latency (no key)
POST /v1/quote/{id}Validate input and price a call; free
POST /v1/run/{id}Run and bill the wallet
GET /v1/requests, /v1/requests/{id}, /v1/results/{id}History (cursor), status/poll, free re-read
POST /v1/keys, POST /v1/agent/signupNew wallet; agent trial key (capped, 7 days)
GET/POST /v1/api-keys, POST/DELETE /v1/api-keys/{id}, …/rotate, …/usageKeys per wallet: limits (daily/weekly/monthly/lifetime), expiry, access level, disable, rotate, usage
GET /v1/me, GET /v1/balanceCurrent key and wallet balance
POST /v1/waitlistGet one email when self-serve top-ups open (no key)
GET /v1/billing, POST /v1/billing/checkout, /v1/billing/auto-topupWallet activity, top-up (returns 501 with a waitlist link until checkout opens), auto top-up
GET /v1/audit-log, /v1/account/settings, POST /v1/feedbackKey/settings change log (JSON/CSV), result retention, bug reports
GET /v1/networkPublic 30-day usage, success, latency per endpoint

Paid calls accept Idempotency-Key (24 h replay, 409 on a different body or while still running). Rate limit: 120 requests/minute per key, X-RateLimit-* and Retry-After headers.

MCP

POST /mcp speaks JSON-RPC 2.0 (initialize, tools/list, tools/call, ping); send Accept: text/event-stream for streamable HTTP. Tools: list_endpoints, search_endpoints, get_endpoint, quote_endpoint, run_endpoint, get_request, get_balance, report_bug, send_feedback. Billing is identical to REST. See MCP.

curl -s -X POST $BASE/mcp -H "Authorization: Bearer $KEY" -H "Content-Type: application/json" \
  -d '{"jsonrpc":"2.0","id":1,"method":"tools/call","params":{"name":"run_endpoint","arguments":{"endpoint":"seo.audit","input":{"domain":"yoursite.com"}}}}'

Errors

Every error: {ok:false, request_id, error:{code, message, retryable, billed:false}}. Full table: errors. Every response carries X-Request-Id.

Crawler etiquette

The crawler identifies as SiteAuditBot, honours robots.txt for every page after the homepage (404 robots = allowed, 5xx = disallowed), fetches at most a few dozen URLs per audit, and only audits sites you own or have permission to audit (see Acceptable Use). Competitors are touched only via robots.txt, sitemap and homepage.

Operator: manual billing owner only

Before a payment processor is connected, sell credit manually: take payment by invoice, then credit the customer's key. Enabled only when ANYAPI_ADMIN_TOKEN is set; otherwise these routes return 404. Not listed in OpenAPI.

ADMIN="X-Admin-Token: $ANYAPI_ADMIN_TOKEN"
# new customer key with $20 credit
curl -s -X POST $BASE/v1/admin/keys -H "$ADMIN" -H "Content-Type: application/json" \
  -d '{"label":"Agency X <ops@agency.test>","credit_usd":20,"reference":"invoice-001"}'
# top up / adjust (kinds: payment, adjustment, refund, trial)
curl -s -X POST $BASE/v1/admin/keys/1/credits -H "$ADMIN" -H "Content-Type: application/json" \
  -d '{"amount_usd":10,"reference":"invoice-002"}'
curl -s $BASE/v1/admin/keys -H "$ADMIN"       # balances, spend, paid
curl -s $BASE/v1/admin/overview -H "$ADMIN"   # totals by credit kind and endpoint
curl -s $BASE/v1/admin/credits -H "$ADMIN"    # credit ledger

Set ANYAPI_PUBLIC_SIGNUP=false to stop self-serve trial keys.