AI Visibility API for ChatGPT, Perplexity, Copilot and Gemini

# ChatGPT (8 credits)
curl "https://api.searlo.tech/api/v1/search/ai/chatgpt?q=best+crm+for+a+small+business&gl=us&hl=en" \
  -H "x-api-key: YOUR_API_KEY"

# The same prompt on Perplexity (2 credits): only the engine changes
curl "https://api.searlo.tech/api/v1/search/ai/perplexity?q=best+crm+for+a+small+business&gl=us&hl=en" \
  -H "x-api-key: YOUR_API_KEY"

import requests

API = "https://api.searlo.tech/api/v1/search/ai"
HEADERS = {"x-api-key": "YOUR_API_KEY"}
params = {"q": "best crm for a small business", "gl": "us", "hl": "en"}

for engine in ["chatgpt", "perplexity", "copilot", "gemini"]:
    resp = requests.get(f"{API}/{engine}", headers=HEADERS, params=params, timeout=180)
    resp.raise_for_status()
    data = resp.json()
    print(f"{engine:<11} {len(data['citations'])} citations  {data['sources']}")

const API = "https://api.searlo.tech/api/v1/search/ai";
const params = new URLSearchParams({ q: "best crm for a small business", gl: "us", hl: "en" });

for (const engine of ["chatgpt", "perplexity", "copilot", "gemini"]) {
  const resp = await fetch(`${API}/${engine}?${params}`, {
    headers: { "x-api-key": "YOUR_API_KEY" },
    signal: AbortSignal.timeout(180_000), // live answers typically take 10-30 s
  });
  if (!resp.ok) throw new Error(`${engine}: HTTP ${resp.status}`);
  const data = await resp.json();
  console.log(engine, data.citations.length, "citations", data.sources);
}

{
  "success": true,
  "searchParameters": {
    "q": "best crm for a small business",
    "type": "ai",
    "engine": "chatgpt",
    "mode": "ui",
    "gl": "us",
    "hl": "en"
  },
  "answer": "Popular choices for small businesses include HubSpot CRM, Zoho CRM and Pipedrive. The right one depends on team size, budget and how much of the sales process you want to automate…",
  "answerMarkdown": "Popular choices for small businesses include HubSpot CRM, Zoho CRM and Pipedrive. The right one depends on team size, budget and how much of the sales process you want to automate…",
  "citations": [
    { "index": 1, "url": "https://www.pcmag.com/picks/…", "title": "The Best CRM Software", "domain": "pcmag.com" },
    { "index": 2, "url": "https://www.forbes.com/advisor/…", "title": "Best CRM Software For Small Business", "domain": "forbes.com" },
    { "index": 3, "url": "https://www.reddit.com/r/smallbusiness/…", "title": "What CRM do you use?", "domain": "reddit.com" }
  ],
  "sources": ["pcmag.com", "forbes.com", "reddit.com"],
  "related": [],
  "model": "",
  "threadUrl": null,
  "complete": true,
  "fetchedAt": "2026-09-29T09:41:12.087Z",
  "source": "searlo.tech(chatgpt)",
  "cached": false
}

share_of_voice.py

"""Share of voice across four AI engines, from the sources field."""
import time
from collections import Counter

import requests

API = "https://api.searlo.tech/api/v1/search/ai"
HEADERS = {"x-api-key": "YOUR_API_KEY"}

ENGINES = ["chatgpt", "perplexity", "copilot", "gemini"]
PROMPTS = [
    "best crm for a small business",
    "easiest crm to set up without an admin",
    "crm with the best free plan",
]
BRAND, DOMAIN = "YourBrand", "yourbrand.com"  # replace with your own
SAMPLES = 3  # answers vary between runs, so sample each prompt


def ask(engine, prompt, attempts=3):
    for _ in range(attempts):
        r = requests.get(
            f"{API}/{engine}",
            headers=HEADERS,
            params={"q": prompt, "gl": "us", "hl": "en", "cache": "false"},
            timeout=180,  # a live answer typically takes 10-30 s
        )
        if r.status_code in (429, 503):  # refused or failed: refunded
            time.sleep(int(r.headers.get("retry-after", 30)))
            continue
        r.raise_for_status()
        return r.json()
    return None


def cites(sources, domain):
    # sources holds hostnames without "www.", so count subdomains too
    return any(d == domain or d.endswith("." + domain) for d in sources)


for engine in ENGINES:
    answers = cited = named = 0
    domains = Counter()
    for prompt in PROMPTS:
        for _ in range(SAMPLES):  # one after another, never in parallel
            data = ask(engine, prompt)
            if data is None:
                continue
            answers += 1
            domains.update(data["sources"])  # each domain once per answer
            cited += cites(data["sources"], DOMAIN)
            named += BRAND.lower() in data["answer"].lower()
    if not answers:
        continue
    print()
    print(f"{engine}: {answers} answers")
    print(f"  {DOMAIN} cited in {cited / answers:.0%}, {BRAND} named in {named / answers:.0%}")
    for domain, n in domains.most_common(5):
        print(f"  {domain:<28}{n / answers:>5.0%}")