Scrapeless
Scrapeless is a web infrastructure platform for AI agents and data pipelines. One API delivers cloud browsers, anti-bot bypass, recursive crawling, residential proxies and AI search-engine data for automation, scraping and data engineering teams.
What is Scrapeless?
Scrapeless presents itself as a single web-infrastructure platform for the age of AI agents: One platform. One API. Five products sit behind that API. Agent Browser is a cloud browser built on a proprietary engine rather than an off-the-shelf headless build, reached through one CDP WebSocket endpoint, with proxy, session TTL, fingerprint and extension settings passed as query parameters. AI Scrapers query seven answer engines (ChatGPT, Perplexity, Gemini, Copilot, Google AI Mode, Google AI Overview and Grok) and return structured fields such as prompt, result_text, model, web_search, links, search_result, content_references, products and ads, mainly for generative engine optimisation and brand-visibility tracking. Web Unlocker, the Universal Scraping API, turns a URL into rendered content. Crawl walks an entire site and returns structured data. Proxy Solutions covers residential, ISP, datacentre and IPv6 pools. Differentiation is claimed at engine level: per-session TLS/JA3 spoofing, Canvas noise, realistic WebGL, randomised audio, profile-locked fonts, timezone and locale consistency, WebRTC leak blocking and Chrome-conformant HTTP/2 frame ordering. Persistent profiles are stored in the cloud, encrypted at rest and isolated per workspace, so an identity survives between runs. MFA Signaling fires a webhook when a session meets two-factor authentication, a human enters the code and automation resumes. Live View streams the DOM in under 100 ms and Session Replay reconstructs a run frame by frame. The platform targets teams that already write code: Puppeteer, Playwright, Selenium, Stagehand, Browser-Use, Crawl4AI, LangChain, CrewAI and n8n are supported, a native MCP server serves Claude Code, Cursor, Cline, Windsurf, Continue and Codex, and SDKs exist for Python, Node, Go and TypeScript. Published figures include 90M+ residential IPs (70M+ on the Agent Browser page), 195+ countries, 99.9% uptime, sub-two-second cold starts, 12+ datacentres, 98%+ challenge resolution and 50+ concurrent sessions as standard, with publisher-displayed ratings of G2 4.8, Trustpilot 4.5, Slashdot 4.8 (4.5 elsewhere on the site) and Tekpon 8.5. The publisher is NST LABS TECH LTD.; the domain was registered on 13 June 2024 and first archived on 11 August 2024. A sixth product, AI Agent, offering no-code agents and an announced marketplace, is waitlist only, not in production.
What it does
- Launch thousands of isolated cloud browser sessions in parallel
- Connect Puppeteer, Playwright, Selenium or Stagehand to a CDP WebSocket endpoint without rewriting existing code
- Solve Cloudflare Turnstile, reCAPTCHA, DataDome and AWS WAF challenges inside the session
- Send a URL and receive the rendered content as HTML, JSON, Markdown or a screenshot
- Crawl an entire website recursively to feed an AI pipeline
- Query seven AI answer engines and retrieve answers, citations and rankings as JSON
- Route traffic through residential, ISP, datacentre or IPv6 proxies with geographic targeting
When to use Scrapeless / When not to
A quick filter to help you decide if Scrapeless is the right fit.
When to use Scrapeless
- Engineering teams running autonomous AI agents in production against real, bot-defended websites
- Data teams collecting at scale behind Cloudflare, DataDome, reCAPTCHA or AWS WAF
- Developers with existing Puppeteer, Playwright, Selenium or Stagehand code who want a cloud endpoint instead of a rewrite
- Marketing and SEO teams measuring generative engine optimisation, tracking what ChatGPT, Perplexity, Gemini, Copilot and Grok say about a brand
- E-commerce, pricing and research teams gathering multi-retailer catalogues, vertical feeds such as airfares, marketplaces and academic journals, or multi-million-page training corpora
When not to use Scrapeless
- Anyone who needs data behind a login, a registration wall or a paywall: the Acceptable Use Policy prohibits collecting non-public information
- Projects touching sensitive or special-category data, health data or children's data, and targets the publisher proactively blocks, namely adult content, government websites and harmful domains
- Proxy resellers and operators of streaming, crypto and NFT, gambling, fake-account, click-fraud or SEO-manipulation schemes, all listed as prohibited activities
- Non-developers looking for a turnkey product: there is no mobile app and no browser extension, and the no-code AI Agent is still a waitlist
- Under-18 users, and buyers who want to compare prices without JavaScript or without an account, since the pricing grid is not rendered in plain HTML
How to use Scrapeless
A typical end-to-end flow, from setup to results.
- Try a prompt in the playground embedded in the AI Scraper product page, with no installation and no account
- Create an account with an email address, a Google account or a GitHub account; the free tier asks for no credit card
- Collect an API key from the dashboard at app.scrapeless.com
- Browser route: replace the local launch with puppeteer.connect() or chromium.connectOverCDP() pointed at wss://browser.scrapeless.com/api/v2/browser, passing token, sessionTTL and proxyCountry in the query string
- API route: make a single call carrying a URL and receive JSON, Markdown, HTML or a screenshot, with webhook delivery available
- MCP route: add a JSON block to the client configuration (npx -y scrapeless-mcp-server with the SCRAPELESS_API_KEY variable), advertised as a 30-second setup
- No-code route: use the native n8n nodes, or Make, Pipedream, Activepieces and Dify
- Install an official SDK for Python, Node, Go or TypeScript if a typed client is preferred
- For AI Scrapers, submit work as asynchronous tasks and collect results pushed to a webhook
- Debug with Live View to watch a session in real time, and Session Replay to replay it frame by frame
Pros & Cons
Pros
- One API covers cloud browser, unblocking, crawling, proxies and AI-engine data, removing the need to assemble several vendors
- Migration from Puppeteer or Playwright is close to trivial: a single connection line changes
- CAPTCHA solving and residential proxies come inside the session, CAPTCHA is not billed separately, and the publisher bills only successful requests
- A permanent free tier with no credit card (one concurrent session and one browser hour per month) sits above genuine usage-based pricing
- Observability that is rare in this market: real-time Live View and frame-by-frame Session Replay, plus MFA Signaling, a capability the publisher presents as missing from most cloud browsers
- A DPA published as a PDF with a named list of subprocessors and Standard Contractual Clauses, claimed SOC 2 Type II with GDPR, CCPA and HIPAA-aligned handling, and an explicit, detailed Acceptable Use Policy
- A broad integration surface (n8n, Make, Pipedream, Activepieces, Dify, LangChain, CrewAI, MCP) backed by documentation, SDKs in four languages and a public status page
Cons
- No postal address is published anywhere: no legal notice, no contact page and no about page across the 1,592 URLs of the sitemap
- The terms of service contain no governing-law clause and no jurisdiction clause, even though section 9.4 of the DPA refers to them explicitly, so the reference points nowhere
- The publisher's country cannot be determined: the site never states it
- A single human channel, market@scrapeless.com, serves support, legal matters and disputes at the same time
- The pricing grid is invisible without JavaScript and the amounts are only served by an internal API, while paid fees and credits are non-refundable except at the publisher's discretion, unused balance is not carried over to the next month, and rates can be increased at any time without notice
- Nothing is published about where customer data is hosted, about model training on customer data or an opt-out, and no retention period is quantified beyond an open-ended link to the life of the account
- Figures contradict each other between pages (90M+ residential IPs on the home page against 70M+ on the Agent Browser page; the Slashdot rating shown as 4.8 and as 4.5), the heavily promoted no-code AI Agent is only a waitlist, and merely using the service grants the publisher the right to use the customer's name and logo for promotion
Pricing & Plans
Scrapeless offers a permanent free plan. The Basic tier costs 0 USD per month with no credit card required, and provides one concurrent session, one browser hour per month and community support. The lowest paid entry point is Growth at 49.00 USD per month, or 529 USD per year. Higher tiers are Scale at 199 USD, Business at 399 USD, Enterprise at 699 USD and Enterprise Plus at 999 USD per month, or 2,148, 4,308, 7,548 and 10,788 USD per year, an annual commitment adding a further 10% discount. It should be noted that a subscription does not buy a quota but a discount on consumption: 10% on Growth, 15% on Scale, 20% on Business, 25% on Enterprise and 30% on Enterprise Plus. Consumption at Basic list rates is charged at 0.09 USD per browser hour, 1.80 USD per GB of residential proxy, 2.50 USD per GB of ISP proxy, 0.80 USD per GB of datacentre proxy, 0.40 USD per GB of IPv6, 1 USD per 1,000 Web Unlocker requests, 3 USD per 1,000 Crawl requests and 1.80 USD per 1,000 AI Scraper requests, with CAPTCHA solving included at no charge. Amounts paid are pre-deposited as a balance and consumed in compute units under a tier-specific CPM algorithm; any balance unused within the month is not carried over, and overage is allowed up to the credit limit before being charged to the card on file. All Stripe payment methods are accepted (card, PayPal, Alipay, bank transfer) together with cryptocurrencies. KYC is not required for a basic plan but is required to raise thread limits or to sign an enterprise plan, through a video call and verification. These figures were read from the publisher's own public pricing API, as the pricing page renders no amount without JavaScript.
- no discount
- consumption charged at list rates
- one concurrent session and one browser hour per month
- no credit card
- community support
- 10% discount on consumption
- technical support in addition to community assistance
- 15% discount on consumption
- 20% discount on consumption
- 25% discount on consumption
- enterprise category
- 30% discount on consumption
- enterprise category
- tailored solutions
- high concurrency
- data cleaning and transformation
- real-time push
- enterprise SLA and SSO
Data, GDPR & hosting
A consolidated view of how Scrapeless handles your data.
GDPR overview
GDPR implementation is documented, not merely claimed. A Data Processing Agreement, version 1.0 effective 7 August 2026, is published as a PDF and covers EU Regulation 2016/679, the UK GDPR and CCPA/CPRA; the customer is Controller, Scrapeless is Processor. Transfers to countries without an adequacy decision rely on Standard Contractual Clauses. Customers may audit once a year on written notice, replaceable by security reports or certifications. The privacy policy lists nine rights: confirmation, access, rectification, erasure, restriction, portability, objection, freedom from automated decision-making and withdrawal of consent, answered within 15 working days, while the Check Your Data page promises a report within 24 hours. Breaches are notified without undue delay and data is returned or securely deleted at contract end. Two gaps: no Article 27 EU representative and no named DPO. Contacts are privacy@scrapeless.com and market@scrapeless.com.
Who owns the data?
Under the Data Processing Agreement the customer is the Controller and Scrapeless the Processor: Scrapeless handles personal data only on the customer's documented instructions, and the customer alone decides which pages are collected and warrants a valid legal basis for the personal data they contain. The website material itself belongs to Scrapeless and its licensors; republishing, reselling, reproducing or redistributing it is prohibited. Comments posted by a user grant Scrapeless a non-exclusive licence to use, reproduce and edit them. Using the service also grants Scrapeless the right to use the customer's name and logo for promotional purposes on its website and materials.
Reuse rights
Scraped output is the customer's to use without asking Scrapeless for permission: the platform only executes documented instructions, and the scope of what is collected, including any personal data inside the pages, is determined by the customer alone, who warrants a lawful basis and remains bound by the Acceptable Use Policy. The material of the Scrapeless website itself may not be reused, republished or resold. On its own side, the publisher declares seven categories of collected data, each tied to one purpose: email address for registration and notifications, password for authentication, payment data for contractual obligations, logs and usage data (IP address, consumption, domains visited, API endpoints called) for service management and improvement, communication logs for exchange quality, API usage data (frequency, parameters, response times) for API optimisation and billing, and browser fingerprint (type, version, system) for the fidelity of browser simulation. Model training appears in none of these purposes. Additional data may come from third parties, notably payment platforms (amounts, history, invoices). Two subprocessors are disclosed, Intercom for email and Airwallex US, LLC for payment. Cookies serve login, do-not-track signals are respected, authentication tokens are kept 24 hours and deleted at logout, and the publisher states it does not sell personal data.
Data retention & training
Hosting summary
Scrapeless publishes no hosting location for customer data. Neither the website, the privacy policy nor the Data Processing Agreement names a country, a region or a cloud provider for the systems that store it, so the jurisdiction of storage cannot be established from first-party sources. The only geographic statements concern the product rather than the data: the proxy network spans 195+ countries and more than twelve datacentres, and the two disclosed subprocessors are United States entities, Airwallex US, LLC for payments and Intercom for email. On security, browser profiles are stated to be encrypted at rest and isolated per workspace. For transfers of EEA or UK personal data to countries without an adequacy decision, the DPA commits to a valid transfer mechanism including Standard Contractual Clauses, with the customer as exporter and Scrapeless as importer. The website itself sits behind Cloudflare on an anycast IP address, which reveals a CDN edge node and not where anything is hosted. Buyers with data-residency requirements should ask for the hosting location in writing before contracting.
Things to keep in mind
Risks and trade-offs to weigh before adopting Scrapeless.
- The publisher's identity is thin: no postal address is published and no first-party source establishes the company's country, so it is not possible to know in advance who the contract is with, or where
- The terms carry neither a governing-law clause nor a jurisdiction clause, and section 9.4 of the DPA refers back to a law the main agreement never designates, so no court is identified in the event of a dispute
- Dispute resolution begins with an email to market@scrapeless.com, the same address used for support and legal matters; the market@ prefix suggests a marketing inbox and is worth testing before relying on it for a formal notice
- The corporate name shown on the site, NST LABS TECH LTD., is close to but not identical with HONG KONG NST LABS TECH CO., LIMITED, the entity the RIPE registry lists as holder of the network's ASN; whether these are the same legal person is not established
- Money terms lean towards the publisher: fees and credits are non-refundable, unused monthly balance is not carried over, rates can be increased at any time without notice, and using the service grants the right to use the customer's name and logo for promotion
- Data governance leaves blanks: no quantified retention period, no published hosting country, and no stated position on whether customer data is used to train models or on how to opt out
- The service is designed to get past anti-bot defences (Cloudflare, DataDome, reCAPTCHA, AWS WAF) using residential proxies, and the contract places responsibility for lawfulness entirely on the customer; the Acceptable Use Policy also forbids collecting data behind a login or a paywall while the Agent Browser page markets authenticated workflows with MFA, and the two are never reconciled, just as network figures differ between pages (90M+ against 70M+ residential IPs)
Setup & Integrations
Technical difficulty
Low to moderate, depending on the route. The fastest is MCP: one JSON block in the client configuration, advertised as a 30-second setup. Developers with existing Puppeteer or Playwright scripts change a single connection line; the API route is one HTTP call carrying a URL; non-developers use native n8n nodes, Make, Pipedream, Activepieces or Dify. Prerequisites are only an account (email, Google or GitHub) and an API key: no infrastructure to provision, no browsers to maintain, with official SDKs in Python, Node, Go and TypeScript. The dashboard and the CU/CPM billing model take longer to master, and code samples require login.
Deployment
Integrations
Supported languages
Behind Scrapeless
Social
Resources
All the official URLs gathered for verification and reference.
Alternatives
Tools that compete with or complement Scrapeless.
Frequently asked questions
What is Scrapeless?
Do I have to rewrite my Puppeteer or Playwright code?
How much does Scrapeless cost?
Is there a free trial?
Is CAPTCHA solving billed separately?
Which bot protections does it get past?
Which AI engines do the AI Scrapers cover?
Is Scrapeless GDPR compliant, and is a DPA available?
Can I scrape data behind a login?
How do I contact Scrapeless, and are refunds possible?
Should you pick Scrapeless?
Scrapeless is a dense, coherent offer for teams that already write code. Five products, Agent Browser, AI Scrapers, Web Unlocker, Crawl and Proxy Solutions, sit behind one API, and the migration cost from an existing Puppeteer or Playwright script is close to zero: one connection line. Observability sits above the market average, with Live View streaming a session live, Session Replay reconstructing it frame by frame, and MFA Signaling handing control back to a human when a run meets two-factor authentication. Economically, the free tier is real and needs no card, and the paid entry point is 49 USD per month, but a subscription buys a discount rather than a quota: the real bill is consumption. Budget on browser hours and proxy gigabytes rather than the monthly fee, and note that unused balance does not roll over and that rates can change without notice. Contractually the picture is split. On one side, a DPA published as a PDF with Standard Contractual Clauses and named subprocessors, plus a detailed Acceptable Use Policy. On the other, no postal address, no governing law and no jurisdiction anywhere in the terms, a single mailbox for support, legal matters and disputes, and nothing on where customer data is hosted. One point deserves attention before signing: this is a service for getting past anti-bot defences, and the contract places responsibility for the lawfulness of what is collected entirely on the customer. The Acceptable Use Policy sets strict limits, notably no data behind a login or a paywall and no sensitive or children's data, worth reading against the intended use, especially as the product pages market authenticated workflows. For technical teams it is a serious candidate. Non-developers will reach it through n8n or Make, and the no-code AI Agent product remains a waitlist.
- Choosing a selection results in a full page refresh.
- Opens in a new window.