Docs › Guides › Choosing the Right Endpoint

Choosing the Right Endpoint

FourA offers four request endpoints, each optimized for a different scenario. Choosing the right one saves time, reduces cost, and improves success rates.

Quick Decision Guide

Use the auto endpoint when:

  • You're targeting a new site and you don't yet know what it needs
  • You want one call that handles direct, proxy-rotated, and browser fallbacks for you
  • You want a session you can replay cheaply on the next call to the same host

Use the single endpoint when:

  • The page is server-rendered (no JavaScript required)
  • You need maximum speed (typically under 1 second)
  • You're hitting APIs or static HTML pages from a host you already know works

Use the browser endpoint when:

  • The page relies on JavaScript to render content
  • Content loads after the initial page load
  • You need the fully rendered DOM

Use the proxy endpoint when:

  • The target site actively blocks requests
  • You need to rotate through multiple IP addresses
  • Previous attempts returned 403 or captcha pages

Endpoint Comparison

Auto (POST /api/auto/)

The smart-fetch endpoint. You pass a URL and (ideally) a validate rule, and FourA walks a cost-aware ladder: cheap direct probe, rotated proxy, full browser. The first rung that returns a response matching your validate wins. On repeat calls to the same host, a warm session is replayed, so the second call is cheap.

curl -X POST https://eu.api.foura.ai/api/auto/ \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com/product/42",
    "validate": {"data": {"accept": ["Add to cart"]}}
  }'

Typical response time: 200ms (warm) to 30s+ (cold solve on a hard site) Best for: New targets, mixed-protection sites, "I just want the page"

For a deeper walkthrough, see the Smart Fetch guide.

Single (POST /api/single/)

The fastest option. Sends an HTTP request with realistic browser-like wire characteristics, without spinning up a browser process.

curl -X POST https://eu.api.foura.ai/api/single/ \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"method": "GET", "url": "https://example.com/api/products"}'

Typical response time: 200ms to 2s Best for: APIs, news sites, blogs, static product pages

Browser (POST /api/browser/)

Opens your URL in a Chrome browser instance. The page loads completely, JavaScript executes, and you get the final rendered HTML.

curl -X POST https://eu.api.foura.ai/api/browser/ \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "url": "https://example.com/spa-app",
    "timeout_ms": 15000,
    "checkText": "data-table"
  }'

Typical response time: 2s to 10s Best for: Single-page apps (SPAs), sites with lazy loading, JavaScript-rendered content

Proxy (POST /api/proxy/)

Combines HTTP requests with automatic proxy rotation. If the first attempt fails or gets blocked, FourA retries through different proxies.

curl -X POST https://eu.api.foura.ai/api/proxy/ \
  -H "X-API-Key: YOUR_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "maxTries": 5,
    "request": {
      "method": "GET",
      "url": "https://example.com/pricing"
    }
  }'

Typical response time: 1s to 5s Best for: E-commerce price monitoring, travel aggregation, sites with bot detection

Auto vs Manual

When should you let auto pick, and when should you call Single, Proxy, or Browser yourself?

Pick auto Pick manual
You don't know what the site needs yet You know exactly which engine the target wants
You want one call that just works You're optimizing the request shape for a known target
You're fine with auto reusing a session it learned You want full control over per-call retry, timeout, and proxy choice
You're paying for a few seconds of probing on the first call Latency on the first call matters more than discovery

Auto isn't always the cheaper pick. If you already know a target works with Single and unblocker on, calling Single directly skips the probe and costs 2 credits. Auto on the same target costs whatever its ladder spends.

When to Combine Approaches

Some workflows benefit from using multiple endpoints:

  1. Discover with auto: pass a validate rule and let the ladder figure out which rung the site needs.
  2. Replay with single: take the session.proxy, session.cookies, and session.userAgent auto returned, then call Single with them for follow-up pages on the same host.
  3. Fall back to browser: if single starts failing, switch to browser rendering.
  4. Add proxy: if you're getting blocked (403 / captcha) without auto, wrap your request in the proxy endpoint for automatic rotation.

This progressive approach keeps cost low while keeping success rates high.

Performance Tips

  • Pass a validate.data.accept substring on protected targets. Without it, auto can't tell a real page from a challenge interstitial.
  • Use the single endpoint by default for known-working hosts and only upgrade when needed.
  • Set checkText in browser requests to avoid waiting for unnecessary content.
  • Set maxTries in proxy requests to control retry behavior (default is 5, max is 90).
  • Keep timeout_ms reasonable: 10 to 15 seconds for most pages, 30s+ for cold auto runs against protected sites.

Next Steps

Last updated: June 30, 2026
Handling Anti-Bot Protection →