GEO Visibility API

Find out whether AI assistants proactively recommend your brand — and whether they get your facts right.

Test version · x402 runs on Base Sepolia testnet only · no real money is charged

What you get

We ask several large language models realistic buyer questions without naming your brand, then audit each answer with a strict judge:

mentioned — did the model name your brand on its own? (A similarly named company does not count.)
rank — your position among the brands it listed.
wrong_facts — statements contradicting facts you supply, or confusion with another company.
cited_sources — URLs the answer relied on (when web search is on).
summary — mention rate, average rank, per-model breakdown and a 0–100 score.

Pricing

Endpoint	What	Price
`POST /v1/x402/ask`	1 question × all models	$0.05 USDC
`POST /v1/x402/checkup`	up to 10 questions × all models (10 defaults if omitted)	$1.00 USDC
`POST /v1/check`	prepaid API key, 1 credit per question	contact us

You are only charged when the check succeeds. If every model call fails, or daily capacity is reached, you get an error and no payment is settled.

Pay per call with x402 (for AI agents)

Call the endpoint; you receive 402 Payment Required with a PAYMENT-REQUIRED header. Any x402 client signs a USDC payment and retries automatically:

import { wrapFetchWithPayment, x402Client } from "@x402/fetch";
import { ExactEvmScheme } from "@x402/evm/exact/client";
import { privateKeyToAccount } from "viem/accounts";

const client = new x402Client();
client.register("eip155:*", new ExactEvmScheme(privateKeyToAccount(process.env.PK)));
const pay = wrapFetchWithPayment(fetch, client);

const r = await pay("https://api.181818.shop/v1/x402/ask", {
  method: "POST", headers: { "Content-Type": "application/json" },
  body: JSON.stringify({
    brand: "Acme Jade", industry: "hetian jade jewelry",
    questions: ["Where can I buy certified hetian jade online?"],
    facts: { founded: 2009, city: "Beijing" }
  })
});
console.log(await r.json());

Prepaid API key

curl -X POST https://api.181818.shop/v1/check \
  -H "X-API-Key: geo_xxx" -H "Content-Type: application/json" \
  -d '{"brand":"Acme Jade","industry":"hetian jade jewelry"}'

curl https://api.181818.shop/v1/balance -H "X-API-Key: geo_xxx"

Request body

Field	Type	Notes
brand	string, required	max 80 chars
industry	string, required	used for default questions
questions	string[]	1–10, max 300 chars each; do not include the brand name
aliases	string[]	up to 5 other names
facts	object	known true facts for wrong-fact detection
models	string[]	subset of `/v1/health` models; default all
web_search	boolean	ground answers in live web results when available

Example response (abridged)

{
  "brand": "Acme Jade",
  "summary": { "checks": 4, "mention_rate": 0.25, "avg_rank_when_mentioned": 3,
               "wrong_fact_count": 1, "score": 25,
               "per_model": { "qwen": {"checked":1,"mentioned":0}, "kimi": {"checked":1,"mentioned":1} } },
  "results": [ { "question": "...", "model_label": "kimi", "mentioned": true, "rank": 3,
                 "wrong_facts": ["says founded in Henan"], "cited_sources": ["https://..."] } ],
  "disclaimer": "AI answers vary between runs; results are a sample, not a guarantee."
}

Notes

Results are cached for 24 hours per brand/question/model.
Rate limit: a few requests per minute per IP. Typical latency 20–90 s for a full checkup.
AI answers are non-deterministic; treat results as a sample.