Bulk HTTP Request Runner & API Scraper: Pagination, JSONPath

Bulk HTTP Request Runner: retries, pagination & JSONPath extraction

Run anything from a few to 100,000 HTTP/API requests from a plain list or from one template plus a list of variables. You get clean, structured results: status, the useful response headers, timing, the parsed JSON, values extracted with JSONPath or JMESPath, and HTML converted to Markdown. Retries, per-host rate limits and pagination are handled for you. Your credentials never appear in the output.

Use it for API backfills, enrichment jobs, health checks, polling endpoints, feeding LLM pipelines with Markdown, or any "call this URL for every row in my spreadsheet" task, without writing or hosting code.

Use cases

• Bulk API calls without code: call a REST API for thousands of IDs from a URL template and variables.
• Paginated API exports: follow Link headers, cursors or page/offset until the end, with JSONPath extraction.
• HTML to Markdown for LLMs: fetch pages and get clean Markdown.
• Uptime and status checks: scheduled checks of endpoints with retries and status/latency in the output.
• No-code glue: a reliable HTTP step for Zapier/Make/n8n flows that need retries and rate limits.

Features

• Two ways to define requests: a requests list (method, url, headers, query, body), or a template with {{placeholders}} plus variables. A placeholder that is the whole value keeps its type, so numbers stay numbers in JSON bodies.
• Authentication: Bearer, Basic, or an API-key header. These are entered in secret input fields, which the platform stores encrypted.
• Secrets are never echoed:
  • Values from the authentication fields are removed from URLs, headers, bodies, extracted values and Markdown.
  • Authorization and cookie headers, and headers or query parameters with names like token, key, secret or signature, show as [redacted].
• Politeness built in: overall concurrency (hard cap 20) and requests per second per host (hard cap 10/s).
• Retries: on network errors, 408, 425, 429 and 5xx, up to 5 times. Retry-After is honored (capped at 120 s). Otherwise backoff is exponential with jitter.
• Pagination:
  • linkHeader (GitHub style rel="next")
  • jsonCursor (a cursor value or next URL at a JSONPath)
  • page and offset parameters
  • maxPages up to 1000, with itemsPath to stop on an empty page
• Parsing: automatic JSON detection, extract rules (JSONPath or JMESPath), and an HTML→Markdown option that keeps main or article content and drops nav, scripts and footers. Raw body storage can be turned off, and maxBodyBytes caps the body size (hard cap 10 MB).

Output (one row per request or page)

id, pageIndex, request (method, redacted URL, redacted headers, hasBody), status, ok, finalUrl, durationMs, attempts, headers (content-type, link, retry-after, rate-limit headers, etag, and others), contentType, bodyBytes, bodyTruncated, extracted, json or body, markdown, error.

Output sample

Real output from a test run (GitHub issues with Link-header pagination and JMESPath [].title / [].number, plus example.com converted to Markdown). The json field is shortened here:

[
  {
    "id": "gh-issues",
    "pageIndex": 0,
    "request": {
      "method": "GET",
      "url": "https://api.github.com/repos/apify/crawlee/issues?per_page=3&state=all",
      "headers": {
        "accept": "application/vnd.github+json"
      },
      "hasBody": false
    },
    "status": 200,
    "ok": true,
    "finalUrl": "https://api.github.com/repos/apify/crawlee/issues?per_page=3&state=all",
    "durationMs": 368,
    "attempts": 1,
    "headers": {
      "content-type": "application/json; charset=utf-8",
      "content-length": "2695",
      "content-encoding": "gzip",
      "link": "<https://api.github.com/repositories/66670819/issues?per_page=3&state=all&after=Y3Vyc29yOnYyOpLPAAABoYhLpAjPAAAAAWvFpZs%3D&page=2>; rel=\"next\"",
      "etag": "W/\"63eb12b0f4552d4469152ba3669f405daf6267d6f23ee575fa6d9c7a5bb7a816\"",
      "cache-control": "public, max-age=60, s-maxage=60",
      "x-ratelimit-limit": "60",
      "x-ratelimit-remaining": "51",
      "x-ratelimit-reset": "1793339689"
    },
    "contentType": "application/json; charset=utf-8",
    "bodyBytes": 14343,
    "bodyTruncated": false,
    "extracted": {
      "titles": [
        "fix(http)!: verify TLS certificates by default",
        "test(basic-crawler): pin that minConcurrency tasks keep running under overload",
        "`npx crawlee create` templates are broken"
      ],
      "numbers": [
        4317,
        4316,
        4315
      ]
    },
    "json": "[… 3 issue objects, omitted here …]",
    "body": null,
    "markdown": null,
    "error": null
  },
  {
    "id": "html-md",
    "pageIndex": 0,
    "request": {
      "method": "GET",
      "url": "https://example.com/",
      "headers": {},
      "hasBody": false
    },
    "status": 200,
    "ok": true,
    "finalUrl": "https://example.com/",
    "durationMs": 66,
    "attempts": 1,
    "headers": {
      "content-type": "text/html; charset=utf-8",
      "content-encoding": "br",
      "last-modified": "Thu, 29 Oct 2026 13:15:03 GMT"
    },
    "contentType": "text/html; charset=utf-8",
    "bodyBytes": 577,
    "bodyTruncated": false,
    "extracted": null,
    "json": null,
    "body": null,
    "markdown": "This domain is for use in documentation examples without needing permission. This is not a service; avoid relying on it for testing and monitoring purposes.",
    "error": null
  }
]

Pricing (pay per event)

Event	Price
request: one successful response (2xx/3xx, or a code you list in allowedStatusCodes); each pagination page counts as one	$0.0008

Failures are never charged. Network errors, timeouts, 4xx/5xx after all retries, and invalid URLs are stored with an error but cost nothing. Retries of a request that finally succeeds count as one request. Cost example: 10,000 API lookups via a template, with 2 pages each = 20,000 successful requests × $0.0008 = $16.

Use it from AI agents (MCP), Zapier, Make, n8n or the API

AI agents (Claude, Cursor, VS Code, ChatGPT and other MCP clients). Add Apify's hosted MCP server with just this Actor as a tool:

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

Sign in with Apify (OAuth) on first connect, or add an Authorization: Bearer <APIFY_TOKEN> header. Then ask your agent, for example:

Use steadyline/http-request-runner to fetch these 30 URLs, convert each page to Markdown and return the titles.

The agent sees this Actor's input schema and output fields, so it can call it without extra prompting. Agents that use the general Apify MCP server can also find it with search-actors.

This Actor uses pay-per-event pricing with limited permissions, so it can also be run through Apify's agentic payments (no Apify account needed for the agent). See the Apify docs.

Zapier, Make and n8n. Use the official Apify app: Zapier "Run Actor" action, Make "Run an Actor" module, or the n8n Apify node. Choose steadyline/http-request-runner, paste the input below, then use "Get dataset items" (or the run's dataset) to send results to Google Sheets, Slack, a CRM or a database.

API (any language). Run it and get the results in one call:

curl -X POST "https://api.apify.com/v2/acts/steadyline~http-request-runner/run-sync-get-dataset-items?token=$APIFY_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{"requests": [{"id": "crawlee-issues", "url": "https://api.github.com/repos/apify/crawlee/issues", "query": {"per_page": "5", "state": "all"}, "headers": {"Accept": "application/vnd.github+json"}}, {"id": "example-page", "url": "https://example.com/"}], "pagination": {"type": "linkHeader", "maxPages": 2}, "extract": {"titles": "$[*].title", "numbers": "$[*].number"}}'

Or with the Python client (pip install apify-client):

from apify_client import ApifyClient
client = ApifyClient("<APIFY_TOKEN>")
run = client.actor("steadyline/http-request-runner").call(run_input={"requests": [{"id": "crawlee-issues", "url": "https://api.github.com/repos/apify/crawlee/issues", "query": {"per_page": "5", "state": "all"}, "headers": {"Accept": "application/vnd.github+json"}}, {"id": "example-page", "url": "https://example.com/"}], "pagination": {"type": "linkHeader", "maxPages": 2}, "extract": {"titles": "$[*].title", "numbers": "$[*].number"}})
for item in client.dataset(run["defaultDatasetId"]).iterate_items():
    print(item)

Works well with other Steadyline Data tools

• Website → RAG knowledge base: get URLs from Site Discovery, fetch HTML pages here with outputMarkdown: true and convert PDFs with PDF to Markdown.
• API monitoring: schedule this Actor every hour and screenshot failing status pages with Website Screenshot.
• Related: Site Discovery (sitemaps, robots.txt, llms.txt), PDF & Document to Markdown, Website Screenshot API.

FAQ

Are my API keys safe? Auth fields are secret inputs (encrypted by Apify), and the Actor redacts them from output, including keys echoed back in response bodies.

Does it render JavaScript? No, it is a plain HTTP client. For rendered pages, use a browser-based Actor; for screenshots see Website Screenshot.

What does it cost? $0.0008 per request or page fetched, so $0.80 per 1,000.

Can I run it on a schedule? Yes. Create an Apify Schedule (or a saved Task with a schedule) and connect a webhook or integration to receive results.

Need a feature or found a bug? Open an issue on the Issues tab of this page or email steadylinedata@proton.me. Built and maintained by Steadyline Data.

Acceptable use

This tool sends the requests you define, so you are responsible for them. Only call APIs and websites you are allowed to use. Follow their terms of service, authentication rules and published rate limits. Do not use it for denial of service, load or stress testing of systems you don't own, credential stuffing, or getting around access controls. Concurrency and per-host rate have hard caps to prevent abuse. Don't send personal data you have no lawful basis to process. We may stop runs that break these rules.

Made by Steadyline Data.