HTTP Request API — Bulk GET/POST Calls Without a Browser

HTTP Request API sends a list of HTTP requests (GET, POST, PUT, PATCH, DELETE, HEAD, OPTIONS) to any public URL or REST API and returns one JSON record per request: status code, response headers, body (text or base64), parsed JSON, an optional fragment picked with a jsonPath or a CSS selector, the redirect chain and timings. Retries with backoff, concurrency, rate limiting and an optional Apify proxy are built in, and the same engine answers synchronously as a Standby HTTP API for AI agents, n8n, Make and scripts.

No browser, no crawling: only the URLs you list are fetched. From US$3.50 per 1,000 answered requests; requests that never get an answer are free.

What it is for

• Call a REST API hundreds or thousands of times from one input (IDs, pages, search terms) and get a clean dataset.
• Health-check a list of endpoints: status, latency, redirects, certificate and DNS errors recorded per URL.
• Fetch raw HTML or JSON for a list of pages and pull out just one value (title, a@href, data.items[*].id).
• Give an AI agent or an n8n/Make workflow a single, safe HTTP tool with retries, proxies and size limits instead of raw fetch.
• Replay POST/PUT/PATCH/DELETE calls in bulk (webhooks, form submissions, API mutations) with per-request bodies.

Input

Field	Description
requests	List of URL strings or objects { url, method, headers, body | json | form, label, jsonPath, selector, timeoutSecs }. Per-request fields override the global ones.
method	Default method for entries without one (GET).
headers	Headers added to every request (authorization, accept, API keys).
jsonPath, selector	Global extraction: dotted path for JSON answers (data.items[*].id, results[0].name, ["key.with.dots"]) or CSS selector for HTML (h1, meta[name=description]@content).
timeoutSecs, maxRetries, concurrency, minDelayMs	30 s per attempt, 2 retries with exponential backoff (Retry-After honoured), 5 requests in flight, optional gap between requests for rate-limited APIs.
followRedirects	Follow up to 10 redirects and report the chain (true), or get the 3xx answer itself.
maxBodyChars, includeBody, includeResponseHeaders, parseJson	Control the size and content of each record. Bodies are cut at 200,000 characters by default and flagged truncated; downloads stop at 10 MB.
browserLikeHeaders	Realistic browser User-Agent/Accept/TLS fingerprint (default) or a plain http-request-api/0.1 User-Agent for APIs. Your own headers always win.
proxyConfiguration	Off by default. Choose Apify datacenter proxy to rotate IPs or residential proxy for sites that block datacenter ranges; the Actor never upgrades to residential on its own.

Example: mixed batch

{
  "requests": [
    { "label": "repo", "url": "https://api.github.com/repos/apify/apify-sdk-js", "jsonPath": "stargazers_count" },
    { "label": "create", "url": "https://httpbin.org/post", "method": "POST", "json": { "name": "demo" }, "jsonPath": "json.name" },
    { "label": "form", "url": "https://httpbin.org/patch", "method": "PATCH", "form": { "a": "1" } },
    { "label": "page", "url": "https://example.com/", "selector": "title" },
    { "label": "ping", "url": "https://example.com/", "method": "HEAD" }
  ],
  "headers": { "accept": "application/json" },
  "concurrency": 5,
  "maxRetries": 2
}

Example: the same API call for many IDs

{
  "requests": [
    "https://api.example.com/v1/products/101",
    "https://api.example.com/v1/products/102",
    "https://api.example.com/v1/products/103"
  ],
  "headers": { "authorization": "Bearer YOUR_TOKEN" },
  "jsonPath": "data.price",
  "includeBody": false
}

Output

One record per input entry, in input order (also for failed and invalid entries, so nothing is lost):

{
  "index": 0,
  "label": "repo",
  "url": "https://api.github.com/repos/apify/apify-sdk-js",
  "method": "GET",
  "finalUrl": "https://api.github.com/repos/apify/apify-sdk-js",
  "redirected": false,
  "redirectChain": [],
  "ok": true,
  "sent": true,
  "statusCode": 200,
  "statusText": "OK",
  "contentType": "application/json; charset=utf-8",
  "contentLength": 7013,
  "responseHeaders": { "content-type": "application/json; charset=utf-8", "x-ratelimit-remaining": "58", "...": "..." },
  "body": "{\"id\":120862636,\"name\":\"apify-sdk-js\",\"full_name\":\"apify/apify-sdk-js\", ...",
  "bodyEncoding": "text",
  "bodyBytes": 7013,
  "truncated": false,
  "json": { "id": 120862636, "full_name": "apify/apify-sdk-js", "stargazers_count": 124, "...": "..." },
  "extracted": 124,
  "extractError": null,
  "responseTimeMs": 727,
  "totalTimeMs": 742,
  "attempts": 1,
  "proxy": "none",
  "error": null,
  "requestedAt": "2026-10-21T04:18:21.512Z"
}

A 404 comes back with sent: true, ok: false, error: "HTTP 404 Not Found" and its body; a timeout or DNS failure with sent: false, statusCode: null and the reason in error. Binary answers (images, PDFs) are stored base64-encoded with bodyEncoding: "base64".

Standby HTTP API

Once Standby is enabled for the Actor, call it directly and get the record back in the same HTTP response:

GET https://fabrikit--http-request-api.apify.actor/?url=https://httpbin.org/get&jsonPath=origin&token=YOUR_APIFY_TOKEN
POST https://fabrikit--http-request-api.apify.actor/?token=YOUR_APIFY_TOKEN
     {"url": "https://httpbin.org/post", "method": "POST", "json": {"a": 1}, "headers": {"x-key": "..."}}
POST ... {"requests": ["https://example.com/", "https://httpbin.org/get"], "selector": "title"}   (up to 20 per call)

Add format=raw to relay the upstream status, content type and body as they are (a tiny HTTP relay for tools that cannot set headers or follow redirects themselves). Query and JSON fields take the same names as the Actor input.

Pricing

Event	Price
request-sent — a request that reached a server and got an HTTP status (2xx, 3xx, 4xx or 5xx)	US$3.50 per 1,000

1,000 answered requests cost US$3.50; 10,000 cost US$35. DNS failures, timeouts, refused connections, blocked private targets and invalid input entries are recorded but not charged, and neither are retries: one request is one event however many attempts it took. Platform compute is included; Apify proxy traffic is billed by the platform only if you turn a proxy on.

Compared with alternatives

• Other bulk HTTP request Actors on Apify Store typically list US$5 per 1,000 requests and charge every attempt; this one is US$3.50, charges only answered requests, keeps the dataset in input order and includes jsonPath/CSS extraction, redirect chains and a Standby API.
• curl / fetch in your own script is free but gives you no dataset, scheduling, proxy rotation, retries or size caps; this Actor is for lists, workflows and agents, not for one-off calls.
• Apify's Web Scraper / Cheerio Scraper render or parse whole pages; here you keep the raw answer and pick one fragment, which is what APIs and health checks need.
• Need the readable text of a web page as Markdown instead of the raw body? Use our URL to Markdown API, which shares the same fetch layer.

Integrations

Run it from the Apify API or the JavaScript/Python clients, schedule it, or plug it into n8n, Make, Zapier and LangChain through the Apify integrations. In Standby mode the Actor behaves like a plain HTTPS endpoint, so any tool that can make a GET or POST call (an MCP server, a custom GPT action, a cron job) can use it as its HTTP tool with retries, proxies and safety limits handled for it.

FAQ

Is this legal? The Actor is a generic HTTP client. You are responsible for the requests you send: respect the terms of the APIs and sites you call, their rate limits and the law that applies to you. Do not use it to attack, flood or scan hosts.

Can it reach private or internal hosts? No. Private, loopback, link-local and cloud-metadata addresses, localhost/.local/.internal names, URLs with embedded credentials and common service ports (SSH, mail, databases) are refused, on every redirect hop too. These entries are recorded with error: "Blocked target: ..." and are free.

Do I need a proxy? Not for APIs and most sites: by default requests leave from the Actor's own IP. Turn on Apify datacenter proxy to rotate IPs, or residential proxy (extra platform cost) only for sites that block datacenter ranges. The Actor never switches proxies without you.

How large can a response be? Up to 10 MB is downloaded per request; the stored body is cut at maxBodyChars (200,000 by default) and truncated tells you when. JSON is parsed from the full download, so json and extracted are complete even when the stored body is cut.

What are the limits? Up to 10,000 requests per run, 20 in flight, 10 redirects per request, 120 s timeout per attempt, 5 retries. Keep concurrency and minDelayMs within what the target allows.

Does it run JavaScript? No. It returns what the server sends. For pages that only render in a browser use a browser-based Actor.

Changelog

• 0.1 (October 2026) — first release: batch and Standby modes, 7 methods, JSON/form/raw bodies, jsonPath and CSS extraction, retries with backoff, redirect chain, size caps, SSRF protection, optional Apify proxy.

---

En español

HTTP Request API envía en lote peticiones HTTP (GET, POST, PUT, PATCH, DELETE, HEAD) a cualquier URL o API pública y devuelve un registro por petición: código de estado, cabeceras, cuerpo (texto o base64), JSON parseado, un fragmento extraído con jsonPath o un selector CSS, la cadena de redirecciones y los tiempos. Incluye reintentos con espera exponencial, concurrencia, límite de tamaño, proxy opcional de Apify y un modo Standby para usarlo como endpoint HTTP desde agentes de IA, n8n o Make. Bloquea direcciones privadas e internas. Desde US$3,50 por 1.000 peticiones respondidas; las que no obtienen respuesta no se cobran.