HTTP Request

HTTP Request calls any public API for you and hands back a tidy table: one row per call, with the status, the headers, the decoded body, the time it took and, when it fails, the exact reason. Make a single call or 10,000 in one run. Works with Clay, n8n, Make, Zapier and AI agents.

🌐 What is HTTP Request?

It is an HTTP client hosted on Apify. Instead of writing a script, you describe the calls you want (address, method, headers, body) and the Actor performs them in parallel, retries the ones that fail for a temporary reason, and stores every answer in a dataset.

From there you can download the results as JSON, CSV or Excel, read them from your code, or have each row pushed to a webhook.

✨ What can this HTTP request tool do?

• 📦 Thousands of calls per run: up to 10,000, with up to 50 running in parallel
• 🔢 Results in your order: the first row is your first request, whatever finished first
• 🧰 All the usual methods: GET, POST, PUT, PATCH, DELETE, HEAD and OPTIONS, with a JSON or text payload
• 🗂️ Defaults for the whole list: write an API key once, and let any request replace or drop it
• 🔗 List parameters done right: {"tag": ["a", "b"]} goes out as tag=a&tag=b
• 🔁 Retries with judgement: temporary failures (network trouble, 429, 5xx) get new tries with growing pauses, and Retry-After is honoured. Final answers such as 404 or 401 come back immediately
• 📄 Any kind of answer: JSON becomes structured data, text stays text, files and images arrive as base64
• ↪️ Redirects made visible: each row shows where the call ended up and how many hops it took
• 🚨 Named failures: dns_error, timeout, connection_error, tls_error, proxy_error... so you know what to fix
• 📬 Webhook delivery: rows can be posted to your endpoint one by one, while the run is still going
• 🌍 Proxy on demand: appear in the country you choose or vary the outgoing IP address
• 🔒 Credentials masked: keys, tokens and cookies never show in the dataset, yet the API receives them intact
• 💸 Billing on answers only: a call that gets no HTTP response costs nothing, and neither do retries

💡 What is it used for?

• 🧱 Enrichment in Clay: query an API that Clay has no connector for, once per row, and bring the answer back into the table
• ⚙️ Loops in n8n, Make and Zapier: replace a slow loop of HTTP modules by one Actor run that returns every answer together
• 📈 Uptime and contract checks: put your endpoints on a schedule and watch statusCode and durationMs over time
• 📣 Notifications to several systems: deliver one event to a list of webhooks, each with its own credentials
• 🗺️ Localised data: read the price, stock or ranking an API serves to visitors of a given country
• 🤖 Tools for AI agents: let an agent reach any REST API through a single, predictable tool

🚀 How to send HTTP requests in bulk

1. Click Try for free and open the Input tab.
2. List your calls in Requests. A URL is enough for a GET.
3. Put what the calls have in common (API key, shared parameters) in the section below, once.
4. Press Start.
5. Open the Output tab: there is one row per call, in the order of your list.
6. Export the table, fetch it through the API, or fill in Webhook URL before starting to receive rows live.

For a single call, leave Requests empty and fill in URL.

💰 How much does it cost to send HTTP requests?

The price is per HTTP response received, and it includes platform usage.

Event	Free	Bronze	Silver	Gold
1,000 responses	$0.30	$0.27	$0.24	$0.20
1,000 responses through a proxy	$1.00	$0.90	$0.80	$0.70
Run start	$0.00005

The proxy price is the same for datacenter and residential proxies.

🎁 Never billed:

• extra tries of the same call
• calls that end without an HTTP response (unknown domain, timeout, refused connection, address not allowed)
• entries that could not be sent (invalid URL, run timeout or spending cap reached)

Calls answered	Free plan, no proxy
1	$0.00035
100	$0.03
1,000	$0.30
10,000	$3.00

Every Apify account receives free credits each month, which cover thousands of calls. If you set a maximum cost in the run options, the Actor stops sending when the cap is reached and marks the remaining entries as skipped_charge_limit.

⬇️ Input

Field	Description
Requests	Your list of calls. Each entry is a URL or { url, method, headers, params, body, id }.
URL	Address for a single call, or default address of the entries.
Method	GET by default.
Headers	Sent with every call. An entry can replace one, or remove it with null.
Query parameters	Appended to the address. A list repeats the name.
Body	JSON payload for write calls. A string is accepted through the API.
Parallel calls	1 to 50, 10 by default.
Timeout	1 to 300 seconds per try.
Retries	0 to 10 new tries, 2 by default.
Statuses worth retrying	408, 425, 429, 500, 502, 503, 504 by default.
Follow redirects, Redirect limit	Up to 20 hops, or none at all.
Decode JSON	Structured data, or raw string.
Size limit per response	1 MB by default, 10 MB at most. With a residential proxy: 100 kB.
Mask credentials in results	Enabled by default.
Webhook URL	Receives each row as it is produced.
Proxy	Apify Proxy (country and group) or your own proxy servers.

A list of calls with a shared API key:

{
  "headers": { "X-API-Key": "YOUR_KEY" },
  "requests": [
    { "id": "acme", "url": "https://api.example.com/companies", "params": { "domain": "acme.com" } },
    { "id": "globex", "url": "https://api.example.com/companies", "params": { "domain": "globex.com" } },
    { "id": "new-contact", "url": "https://api.example.com/contacts", "method": "POST", "body": { "email": "jane@acme.com", "tags": ["lead", "2026"] } }
  ],
  "maxConcurrency": 5
}

One call:

{
  "url": "https://jsonplaceholder.typicode.com/todos/1"
}

A call seen from the United States, with rows pushed to a webhook:

{
  "requests": [
    { "id": "sku-1042", "url": "https://shop.example.com/api/products/1042" }
  ],
  "proxy": { "useApifyProxy": true, "apifyProxyCountry": "US" },
  "webhookUrl": "https://your-app.example.com/hooks/http-results"
}

⬆️ Output

Field	Description
id	The id of your entry, or request-1, request-2...
index	Rank of the entry in your list, from 0
url	Address called, parameters included
finalUrl	Address that answered, once redirects were followed
method	HTTP method
statusCode, statusText	Status returned by the server
ok	true when the status is between 200 and 299
requestHeaders	Headers you supplied, credentials masked
responseHeaders	Headers of the answer
body	The answer: structured JSON, text, or base64
bodyEncoding	json, text, base64 or none
bodyBytes	Size of the answer in bytes
truncated	true when the size limit cut the answer
durationMs	Time spent on the call, retries included
attempts	Number of tries
redirectCount	Number of redirects followed
proxyUsed	Whether the call went through a proxy
errorType, errorMessage	Reason, when there is no HTTP response
webhookStatus	What your webhook answered, when one is set
fetchedAt	Date and time of the call

A row for a successful POST:

{
  "id": "create-post",
  "index": 1,
  "url": "https://jsonplaceholder.typicode.com/posts",
  "finalUrl": "https://jsonplaceholder.typicode.com/posts",
  "method": "POST",
  "statusCode": 201,
  "statusText": "Created",
  "ok": true,
  "requestHeaders": { "content-type": "application/json" },
  "responseHeaders": { "content-type": "application/json; charset=utf-8" },
  "body": { "title": "Hello", "userId": 1, "id": 101 },
  "bodyEncoding": "json",
  "bodyBytes": 50,
  "truncated": false,
  "durationMs": 212,
  "attempts": 1,
  "redirectCount": 0,
  "proxyUsed": false,
  "errorType": null,
  "errorMessage": null,
  "fetchedAt": "2026-10-06T21:40:07.115Z"
}

A row for a call that never got an answer:

{
  "id": "old-partner",
  "url": "https://api.partner-that-closed.example/v2/status",
  "statusCode": null,
  "ok": false,
  "attempts": 1,
  "errorType": "dns_error",
  "errorMessage": "getaddrinfo ENOTFOUND api.partner-that-closed.example"
}

errorType	What happened	New try?
dns_error	The domain name does not resolve	Only if the resolver itself was unavailable
timeout	The answer did not arrive in time	Yes
connection_error	The server refused or dropped the connection	Yes
tls_error	The HTTPS certificate is not valid	No
proxy_error	The proxy failed, or your plan has no access to it	Yes, unless access was denied
network_error	Another transport problem	Yes
blocked_url	The address is not on the public internet, or uses a port reserved for another protocol	No
too_many_redirects	The redirect limit was exceeded	No
invalid_request	The entry has no valid URL or an unknown method	No
skipped_timeout	The run was about to time out before this entry	No
skipped_charge_limit	Your spending cap was reached before this entry	No

🧩 How to use HTTP Request with Clay, n8n, Make and Zapier

Any tool able to send a POST can drive the Actor. This endpoint launches a run and replies with the rows once they are ready:

POST https://api.apify.com/v2/acts/data_xplorer~http-request/run-sync-get-dataset-items

• Clay: make one run per table row with url, your headers and an id, then map fields of the returned body to columns.
• n8n, Make, Zapier: turn your items into the requests list, launch one run, then iterate over the rows. Since the order is kept, row N belongs to item N.
• Large lists: fill in Webhook URL and let your scenario be triggered by each row, instead of keeping a connection open until the end.

n8n, Make and Zapier also have an Apify app, if you prefer modules to raw calls.

🔌 Can I use HTTP Request as an API?

Yes, through the Apify API or the Apify clients.

cURL:

curl -X POST "https://api.apify.com/v2/acts/data_xplorer~http-request/run-sync-get-dataset-items" \
  -H "Authorization: Bearer <YOUR_API_TOKEN>" \
  -H "Content-Type: application/json" \
  -d '{"requests":["https://jsonplaceholder.typicode.com/todos/1","https://jsonplaceholder.typicode.com/users/1"]}'

Python:

from apify_client import ApifyClient

client = ApifyClient("<YOUR_API_TOKEN>")
run = client.actor("data_xplorer/http-request").call(run_input={
    "requests": [
        {"id": "todo", "url": "https://jsonplaceholder.typicode.com/todos/1"},
        {"id": "user", "url": "https://jsonplaceholder.typicode.com/users/1"},
    ],
})
for row in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(row["id"], row["statusCode"], row["durationMs"])

Node.js:

import { ApifyClient } from 'apify-client';

const client = new ApifyClient({ token: '<YOUR_API_TOKEN>' });
const run = await client.actor('data_xplorer/http-request').call({
    requests: [
        { id: 'todo', url: 'https://jsonplaceholder.typicode.com/todos/1' },
        { id: 'user', url: 'https://jsonplaceholder.typicode.com/users/1' },
    ],
});
const { items } = await client.dataset(run.defaultDatasetId).listItems();
for (const row of items) console.log(row.id, row.statusCode, row.durationMs);

🤖 Use HTTP Request with AI agents (MCP)

Through the Apify MCP server, assistants such as Claude, Cursor or VS Code can use this Actor as a tool. The agent decides which endpoints to call, and gets back structured rows it can reason on, including a named error when a call fails.

Server URL restricted to this Actor:

https://mcp.apify.com?tools=data_xplorer/http-request

Claude Code:

$claude mcp add --transport http http-request "https://mcp.apify.com?tools=data_xplorer/http-request"

MCP config file of Cursor, VS Code and similar clients:

{
  "mcpServers": {
    "http-request": {
      "url": "https://mcp.apify.com?tools=data_xplorer/http-request"
    }
  }
}

Things you can ask:

• "Fetch these 15 product endpoints and list the ones whose price changed since my last note."
• "Are all the URLs in this sitemap answering with a 200? Give me the broken ones and why."
• "Send this JSON to my staging webhook and show me what it replies."

🛡️ Safety

• Only addresses on the public internet can be called. Loopback, private networks and reserved ranges are refused, and the check is repeated on every redirect and on the IP address the name resolves to.
• When a redirect leaves the original site, your Authorization header and cookies stay behind.
• Results are safe to share: credentials are masked unless you decide otherwise.

❓ FAQ

📦 How many calls fit in one run?

10,000. Rows are written to the dataset in groups while the run is going, so a long list is never held back until the end.

🔢 Is the order of the results guaranteed?

Yes. Calls are made in parallel, but rows are stored in the order of your list. index and id are there as well if you sort or filter later.

🔁 What if the API replies 429 or 503?

Those two are treated as temporary. The Actor pauses, longer after each failure, and follows Retry-After when the API provides it (15 seconds at most). After the last try you receive the answer as it is, billed as one response.

🔑 How do I pass an API key?

In Headers (for instance X-API-Key or Authorization) or in Query parameters, depending on the API. It reaches the API unchanged and is masked in the stored results.

🍪 Can I read the cookies a server sets?

Yes: disable Mask credentials in results and set-cookie appears in responseHeaders.

📎 Can it download a file?

Yes, within the size limit. Its content is in body as base64, and bodyEncoding says so.

🖥️ Does it run JavaScript or sign in to websites?

No. It speaks plain HTTP, which is all an API needs and the reason it is quick and inexpensive. Pages that depend on a browser call for a scraper built for them.

🌍 Which proxy countries can I pick?

The ones offered by the proxy groups of your Apify plan. If a country is not available to you, the row says so with a proxy_error.

🛟 Something is missing or broken. What now?

Tell us in the Issues tab. Every report gets an answer.

🔗 Other tools you may like

• Google Flights Scraper: flight prices, schedules and booking links from Google Flights
• Google Trends Scraper: daily trending searches by country and keyword interest over time
• Google News Scraper: Google News articles in real time, with titles, sources, dates, images and URLs
• TikTok Ads Library FAST: competitor ad campaigns and creatives from the TikTok Ads Library