Skip to content

Rumble Video and Creator Scraper API

Send one Rumble URL and receive rendered, parse-ready data for videos and creators.

OUTPUT FORMATSHTMLJSONMarkdown

JavaScript execution, proxy rotation, and bot-detection handling are managed by the API, so you can focus on the fields you need rather than browser infrastructure.

What to collect from Rumble

Most teams use this workflow for three distinct jobs. Keep them separate when designing a collection, because each job needs different URLs and different fields.

  • Track creator videos: watch selected creators for new uploads, view counts, engagement, and channel growth, and maintain a niche watch list.
  • Monitor video performance: collect views, likes, comments, and watch percentage over time to compare videos and learn what audiences respond to.
  • Identify trending content: review trending and discovery pages to spot topics and creators gaining traction before they become widely visible.

A practical starting checklist: choose one job, collect a small set of representative URLs, decide which fields prove the job is working, test rendered output, then scale only after the field map below matches your needs.

How it works

The sequence below is the shortest reliable path from a blank environment to a parsed Rumble response. Treat each step as a checkpoint: if the next step fails, the problem is usually in the step just completed rather than in the Rumble page itself. The allowance available on a new account is set out in the plan and trial terms published for this service; check those terms before assuming a specific number of free requests. The same applies to any credit or concurrency figure quoted elsewhere on this page — the published terms are the authority, not a remembered number from an earlier version of the product. If you are evaluating the service before committing, a small pilot of a few dozen URLs is usually enough to confirm that the rendered output contains the fields you need. That pilot also reveals whether your target pages need JavaScript rendering, a stealth proxy path, or country targeting, which are the three settings most likely to change your per-request cost. Budget for those settings explicitly rather than discovering them after a large run. The SDKs for Python, Node, Go, PHP, and Ruby wrap the same HTTP contract, so a request that works in one language works in the others with only the client call changed. If you prefer not to install anything, the request builder in the dashboard produces a copy-ready snippet in Python, Node.js, cURL, or the other supported languages; paste it, substitute your own API key, and send it. Keep the API key out of source control and out of any snippet you share, since a leaked key is equivalent to a leaked account. When you move from the pilot to production, log the request parameters alongside each response so that a later change in output can be traced to a specific setting rather than guessed at.

Control and field map

Use this map to decide which control solves which problem. Do not enable everything by default — each control addresses a specific failure.

  • render_js: run JavaScript in a real headless browser and return post-render HTML. Use when content is missing from the initial page.
  • stealth_proxy: use a stealth proxy path for CAPTCHA-heavy or aggressive responses.
  • country_code: route through a residential IP in a specific country when content varies by geography.
  • extract_rules: return structured JSON using CSS or XPath rules for exact elements.
  • ai_extract_rules: describe the data you want in natural language and receive structured JSON.
  • Screenshots: capture a full-page image for visual verification alongside extracted data.
  • JavaScript scenarios: click, scroll, or type before extraction to reveal hidden or dynamically loaded content.
  • Markdown output: receive a clean Markdown version of a video page for LLM-based creator or trend analysis.
  • n8n automation: drive creator-monitoring workflows without writing scraper code.

Fields commonly returned for a Rumble page include video metadata, view count, like count, comment count, upload date, creator handle, and hashtags. In one call, video metadata, view counts, engagement signals, hashtags, and creator information can be parsed together.

Failure distinctions and limits

Three failure modes account for most disappointing Rumble results, and each has a different remedy. Diagnose before you change anything, because enabling every option at once raises cost without telling you which setting actually fixed the problem. The first mode is content that is visible in a normal browser but absent from the response. That usually means the page assembles its data with JavaScript after the initial HTML arrives, so render_js is the relevant setting; it runs the page in a headless browser and returns the post-render HTML. The second mode is an outright block, a CAPTCHA challenge, or an IP-based rejection. Here the request itself is being refused, so stealth_proxy is the relevant setting, working alongside rotating residential proxies so that traffic resembles ordinary browsing rather than a single repeated source. The third mode is content that loads but differs from what you expect — a different regional feed, a missing section, or inconsistent ordering. That points to geography-dependent delivery, and country_code routes the request through a residential IP in the chosen country. Geotargeting reaches a wide set of countries and territories, with a smaller subset available on classic proxies, so confirm that your target country is covered before building a workflow around it. On throughput, the constraint is the number of concurrent requests allowed by your plan rather than a per-target rate limit. If a destination throttles aggressive patterns, space requests out or distribute them across multiple sessions. Credit accounting has its own rules: credits are valid for one month and do not roll over, requests start at one credit each, and JavaScript rendering and premium proxies add per-request cost. The exact multipliers are set by the published pricing terms, so read those before sizing a run. Where team management is included in a plan, it allows multiple users to be invited to an organization with individual credit and concurrency quotas assigned to each.

Pricing and plan selection

Choosing a plan is mostly an exercise in estimating three numbers: monthly request volume, the concurrency your workflow needs at peak, and how many requests will require the costlier options. Prices, included features, and cancellation terms are published on the pricing page and in the applicable service terms; those documents are the authority for any figure quoted in a sales conversation or an older copy of this page. All prices are exclusive of VAT, so add the applicable rate for your jurisdiction when comparing against alternatives. Work through the feature list deliberately rather than assuming everything is bundled. JavaScript rendering, premium proxies, geotargeting, screenshots, and extraction rules each affect what a request costs, and priority email support, a dedicated account manager, and team management are typically reserved for higher tiers. A useful method is to take your pilot run, multiply the request count by the per-request cost of the options you actually used, and compare that against each tier's included allowance and concurrency ceiling. If the estimate lands near a boundary, size up rather than down, because concurrency limits are enforced and a queue that stalls at peak is worse than a slightly larger bill. Before committing to a tier, confirm the trial terms and any free allowance directly with the provider, since those terms can change and should not be inferred from this page.

Compliance, scale, and speed

The service states GDPR and CCPA compliance, and that it does not collect or store personal data from scraped sites unless a user explicitly requests it. Proxy rotation and headless-browser rendering are intended to mimic real browsing behaviour and reduce blocking risk. The platform is described as scaling with thousands of headless browsers and rotating proxies, with results typically delivered in 1–5 seconds and high accuracy across most sites, including JavaScript-heavy ones. These operational claims should be confirmed by the owner before publication as current guarantees.

Frequently asked questions

How do I start scraping Rumble?

Register an account, then locate your API key in the dashboard. With the key in hand, issue a GET request that includes the Rumble URL you want to collect. The service renders the page, rotates proxies, and returns the resulting data for you to parse. Any free allowance or trial credit attached to a new account is governed by the current published terms, so check those terms rather than relying on a figure quoted elsewhere. A sensible first test is a single video page and a single creator page, which confirms both that your key works and that the rendered output contains the fields you intend to collect.

Will I get blocked when scraping Rumble?

Residential proxy rotation and a real headless browser make requests resemble ordinary browsing. When a destination responds aggressively, enable stealth_proxy for that request.

Can I scrape Rumble without writing code?

Yes. A dashboard request builder accepts a URL and parameters, then produces a ready-to-copy snippet in Python, Node, Go, PHP, Ruby, or cURL.

How do I handle JavaScript when scraping Rumble?

Set render_js=true. The page runs in a headless browser on the server, JavaScript executes, and the returned HTML reflects the rendered state.

Can I scrape Rumble from a specific country?

Yes. Supply country_code to route the request through a residential IP in the chosen country, which helps when pages differ by location.

What's the rate limit for scraping Rumble?

Limits are based on concurrent requests for the selected plan rather than a per-target rate limit. If a destination throttles aggressive patterns, space requests out or use multiple sessions.