Apify

  • Apify — Full Profile (English, 2026-09)

    Headquartered in Prague, Czech Republic, Apify Technologies s.r.o. was founded in 2015 by Jan Curn and Jakub Balada, participants of the 2015/2016 YC Fellowship program. The company operates with a VC-independent leaning while retaining formal VC backing. Prior to its first institutional financing, Apify achieved profitability through a bootstrapped operation model. It secured a €2.8M funding round in April 2024 led by J&T Ventures and Reflex Capital, with total capital raised reaching approximately $4.5M per Simplify data. As a YC alumni with limited venture funding, Apify remains unacquired, non-public, non-nonprofit, and is not classified as a Big Tech platform primitive. It falls into the VC-independent vendor bucket alongside n8n, Databox, Typeform and Documenso, clearly separated from pure bootstrapped players including Todoist and QuickChart. Apify provides a stable, official first-party GA MCP server hosted at mcp.apify.com, adopting OAuth streamable-HTTP transport while discontinuing SSE support as of April 2026. Its official open-source MIT-licensed MCP server @apify/actors-mcp-server holds 3.7K to 4.8K GitHub stars, unlocking a massive tool ecosystem of 26K–30K Apify Actors available for agent invocation. Ecosystem-wise, Apify joins the top-tier stable first-party GA server cluster with Klaviyo, Feedly, Slack, Databox and Tavily. It is differentiated from n8n dual-GA architecture, Google Sheets/Gmail first-party preview stack, Typeform beta-grade MCP, Nanonets Composio-exclusive integration model, and the zero-first-party MCP cohort covering Documenso, QuickChart, Semantic Scholar and Zotero.

    Website

    Field Value
    Official Website https://apify.com · docs: docs.apify.com · MCP docs: docs.apify.com/platform/integrations/mcp · REST base: api.apify.com/v2 · MCP host: https://mcp.apify.com · OSS MCP: github.com/apify/apify-mcp-server · SDK: github.com/apify/apify-client-js / apify-client-python · Crawlee: github.com/apify/crawlee (6K★+)
    Company Apify Technologies s.r.o., Prague; founded 2015​ Curn+Balada (YC Fellowship); VC-backed but lightly​ (€2.8M J&T+Reflex Apr 2024, ~$4.5M total, profitable pre-2024); 243 employees (Jul 2026), 26K-30K Actors in Store, 10K-25K customers (T-Mobile/MS/Samsung/Accenture/Siemens/Groupon), 1PB+/mo processed; AGPL-3.0 client/SDK, proprietary platform; SOC 2-aware, GDPR ToS, proxy-compliant tier; Cloud-only managed (no on-prem platform, Crawlee self-hostable)
    Status Active, Apify Store 26,929 Actors (Apr 2026), MCP server GA Jun 2025, x402/Skyfire agentic-pay Jun 2026
    Category ai-productivity-data
    Pipeline Stage none

    Title

    Field Value
    Product Apify (Actor cloud + Store + MCP server)
    Descriptive Title Apify — Web Scraping & Browser-Automation Cloud with 26K+ Prebuilt Actors as Agent Tools: serverless “Actor” units (Node/Python, Playwright/Crawlee) run on managed infra with proxy rotation/CAPTCHA/JS-render/scheduling/storage/webhooks → REST api.apify.com/v2 (actor calls, datasets, key-value store, request queues, runs) + official first-party MCP server​ mcp.apify.com OAuth streamable-HTTP exposing 5K-30K Actors as discoverable tools (rag-web-browser/google-maps-scraper/linkedin-scraper/amazon-product/etc.) to Claude/Cursor/VS Code/ChatGPT + OSS @apify/actors-mcp-server MIT stdio; credit-metered on compute units (CU)​ + per-Actor pricing, Free 29 → Scale 999 → Enterprise custom; 2015 Prague Curn+Balada YC-fellowship €2.8M-2024 vc-independent-leaning MCP-first-party-GA-actor-marketplace
    One-line Positioning Agent says “scrape top 10 Prague cafés from Google Maps” → Claude calls mcp.apify.com apify–google-maps-scraper tool → Actor runs on Apify cloud → structured JSON back into context; billing is platform credit (CU + Actor fee) not per-seat (reverse Motion) not per-interaction (reverse Qualtrics) not endpoint-credit (reverse Tavily) not image-count (reverse QuickChart) not 1-RPS (reverse Semantic Scholar); ownership VC-backed-private (same macro-bucket as n8n/Databox/Typeform/Documenso, distinct from Tavily acquired-kept / Zotero-CDS nonprofit / QuickChart bootstrap).

    Features

    Field Value
    Actor Model Serverless containerized unit (Node.js/Python, Crawlee/Playwright/Puppeteer), input schema JSON, output to dataset (JSON/CSV), key-value store (files), request queue; 26,929 Actors in Store (Apr 2026), Apify-built + community-monetized; popular: google-maps-scraper, linkedin-scraper, amazon-product-scraper, tweet-scraper-v2, instagram-scraper, rag-web-browser, web-fetch, website-content-crawler
    Managed Infra Automatic proxy rotation (datacenter/residential/Google SERP proxies), CAPTCHA solving tier, JS rendering, headless Chrome, scheduling (cron), webhooks on run/fail, storage retention (7-day free → longer paid), scaling concurrency, run log streaming
    SDK / Frameworks Crawlee (Node 2022, Python Jul 2024) 6K★+, apify-client-js/python, LangChain ApifyWrapper/ApifyDatasetLoader, LangGraph node, CrewAI tool, Zapier/Make/n8n connectors, Airflow operator
    MCP (first-party GA) Official​ https://mcp.apify.com (OAuth Bearer or URL param, streamable-HTTP, SSE removed Apr 2026) + OSS npx @apify/actors-mcp-server (APIFY_TOKEN env, Node 18+) MIT ~3.7-4.8K★; tools = dynamic Actor discovery (find_actor/list_actors) + run_actor + actor input-schema introspection + dataset read + kv-store read + resources/read passthrough to api.apify.com/v2 + prompts cap (empty); 30 req/s/user limit; telemetry opt-out flag; x402 + Skyfire + AGI token agentic-pay​ (Jun 2026) lets agents pay per Actor run without Apify account; joins Klaviyo/Databox/Tavily first-party-GA-server cluster
    API REST v2 Bearer token, actor/client/run/dataset/kv/queue/store endpoints, ?token=, pagination, webhooks, 30 req/s MCP / higher on REST paid tier; no GraphQL
    Safety / Legal Proxy-compliant tier (residential opt-in), robots.txt respect toggle per Actor, ToS disclaimer in MCP docs (“respect website terms”), OAuth scopes for MCP connectors (Notion/GitHub/Slack Jun 2026)

    Pricing (2026-09, platform credit = compute units + per-Actor fee, no seats, Free $5/mo)

    Field Value
    Free 5 platform credits/mo, no card, 4GB Actor RAM cap, 7-day data retention, 20 shared datacenter proxies, 9 team seats, 30 req/s MCP
    Starter 0.30/CU reference), higher RAM, 30-day retention, shared proxies, community support
    Scale $199/mo​ — more CU, residential proxy pool, 60-day retention, priority queue, email support
    Business $999/mo​ — large CU bundle, dedicated proxy lanes, 180-day retention, SLA-ish, solutions support
    Enterprise Custom — unlimited CU, VPC/private proxy, SSO/SAML, DPA, custom Actor SLAs, onboarding
    Compute Unit (CU) 1 CU ≈ 1 vCPU-sec + 128MB-sec baseline; Actor run cost = CU consumed × plan rate + Actor-specific per-result surcharge (community Actors set own price, some free); rag-web-browser ~few CU/query, google-maps-scraper ~tens CU/100 results
    Type usage-credit-platform-CU-no-seat; reverse Motion per-seat, reverse Qualtrics per-interaction, reverse Tavily endpoint-credit (Tavily credit≠CU but both credit-metered), reverse Databox AI-credit, reverse QuickChart image-count, reverse Gmail 80M-unit, reverse Semantic Scholar 1-RPS, reverse Zotero storage-quota
    Gotcha Community Actor quality variance (abandoned/broken schemas), CU math opaque until invoice, residential proxy upsell, x402 agentic-pay needs Skyfire/AGI wallet, MCP 30 req/s not 1k

    Reviews

    Field Value
    G2 4.6/5 ~30-40 reviews​ (Apify has a claimed G2 profile, dev-tool B2B pattern, smaller n than Klaviyo 580/Capterra 410 but real) → print G2 4.6/~35 claimed-devtool (solid-er than Tavily 2-review, thinner than Klaviyo); praise proxy/anti-block, gripe CU opacity + community Actor rot
    Capterra 4.7/5 ~20-30 reviews​ (Capterra dev-tool infra) → Capterra 4.7/~25 directional-solid
    Trustpilot No B2C page​ (B2B dev infra) → no-rated-consumer-profile excluded (aligns FFmpeg/QuickChart/Semantic Scholar/Zotero)
    Dev / MCP sentiment apify-mcp-server 3.7-4.8K★, Crawlee 6K★, AIMultiple/Context.dev 2026 “marketplace breadth unmatched, CU opaque, community Actor variance”, Tekpon 4.6 normalized; praise “26K Actors = never write scraper again”, MCP dynamic-tool-discovery best-in-class; gripes CU surprise bills, Actor abandonment, not for single-URL markdown (Firecrawl simpler), x402 early
    Praise Actor marketplace 26K+ covers niche sites no general crawler handles, managed proxy/CAPTCHA/JS-render offload, official MCP GA with dynamic discovery (agents pick Actor at runtime), Crawlee OSS escape hatch, x402 agentic-pay frontier, LangChain/CrewAI native
    Gripes Compute-unit billing opaque (hard to forecast 29 Starter thin for real jobs (scale to $199 fast), residential proxy upsell, single-URL RAG better served by Firecrawl/Tavily, ToS/robots compliance is user’s responsibility

    Review Sources Breakdown

    Field Value
    G2.com 4.6/5 ~35 reviews claimed devtool-B2B → G2-solid-er-than-tavily-thinner-than-klaviyo
    Capterra 4.7/5 ~25 → Capterra-directional-solid
    Trustpilot none → no-rated-consumer-profile excluded
    GitHub apify-mcp-server 3.7-4.8K + Crawlee 6K dev-sentiment directional
    AIMultiple/Context.dev/Tekpon 4.6 norm directional
    Conclusion​ Prints “G2 4.6/~35 + Capterra 4.7/~25 devtool-B2B real (not unclaimed like Tavily 2-review, not absent like Zotero/QuickChart), GitHub MCP 3.7-4.8K, dev consensus marketplace-breadth-unmatched CU-opaque-actor-variance” — aligns vc-independent SaaS G2-rated pattern (Klaviyo/Databox/Typeform), distinct from nonprofit/community-only none-cluster.

    Access

    Field Value
    Web App console.apify.com — Actor Store, runs, datasets, billing, API keys, MCP config
    API REST api.apify.com/v2 Bearer token, actor call /runs /datasets /kv /queues /stores, webhooks, 30 req/s MCP / higher REST
    MCP First-party GA​ — hosted https://mcp.apify.com (OAuth or ?token= or Bearer header, streamable-HTTP, SSE dropped Apr 2026) + local npx @apify/actors-mcp-server (APIFY_TOKEN, Node 18+); tools find/run Actor, read dataset/kv, resources/read passthrough; Claude/Cursor/VS Code/ChatGPT/MCP-inspector; x402/Skyfire/AGI agentic-pay supported
    SDKs apify-client-js, apify-client-python, Crawlee (JS/Py), community Go/Ruby/PHP; LangChain ApifyWrapper, LangGraph, CrewAI, Zapier/Make/n8n
    Login apify.com/login → Settings/API & Integrations → token; OAuth for mcp.apify.com hosted
    Self-host Crawlee + apify-client can run locally against your own browser/proxy, but Actor runtime + Store + datasets + MCP hosted backend are Apify Cloud only​ (no on-prem Apify platform binary, contrast QuickChart/Documenso/Zotero-translation-server)

    Best For

    Field Value
    Audience 1 Teams needing prebuilt scraper for Google Maps/LinkedIn/Amazon/Instagram/TikTok without writing Playwright
    Audience 2 AI-agent builders wanting Claude/Cursor to dynamically discover+run 26K web tools via official MCP (runtime Actor pick)
    Audience 3 Recurring scheduled extraction jobs (price monitoring, SERP tracking, lead scrape) with webhook→warehouse
    Audience 4 Developers wanting Crawlee OSS for custom scrapers but managed proxy/rendering on Apify Cloud for runs
    Audience 5 Autonomous agents using x402/Skyfire to pay-per-Actor-run without human Apify account (2026 frontier)
    Audience 6 Enterprise needing residential proxy + SSO + DPA for compliant collection at scale

    Not For

    Field Value
    Exclude 1 Single-URL→markdown for RAG (one-shot) — Firecrawl/Tavily cheaper/simpler, Apify CU overkill
    Exclude 2 Teams refusing credit-metered opacity — CU+per-Actor math needs invoice literacy (reverse Motion per-seat buyers)
    Exclude 3 On-prem/air-gap scraping platform — Apify Cloud only, Crawlee alone isn’t Apify
    Exclude 4 Buyers needing first-party MCP with write-to-notion/github/slack native — Apify MCP runs Actors, Jun 2026 MCP-connectors add OAuth-to-apps but not same as Databox account-scoped
    Exclude 5 Niche sites with no Store Actor + no dev to write Crawlee — community Actor rot risk
    Exclude 6 Strict-ethical/no-proxy users — Apify’s value is anti-block; if you reject residential proxy, use plain Crawlee/Scrapy

    Competitors

    Field Value
    Firecrawl General /scrape /crawl /map /search → clean markdown+JSON, LLM-native, official MCP, $16/mo hobby, weaker on site-specific Actors (no LinkedIn/Amazon deep extractors), better for RAG-one-shot
    Bright Data Proxy-first (residential 195+ countries), Web Unlocker + 40+ scrapers, $499/mo min/product, no first-party MCP (60+ tools via 3rd-party), raw HTML mostly, enterprise compliance scale
    Oxylabs Proxy + Web Scraper API, $49/mo entry, 175M residential, ai_scraper MCP (separate AI Studio key), results-based, enterprise setup heavier than Apify
    ScrapingBee / ScraperAPI Simple unblocking API, credit-based, no Actor marketplace, you own parsing
    Crawl4AI / Crawlee-self-host OSS self-host zero cost, no proxy/store/managed infra
    Zyte AI Crawler + 93% success on protected, pay-per-success, no 26K Actor marketplace
    Context.dev / KnowledgeSDK One-API markdown+semantic+webhook, no marketplace, lowest engineering overhead

     

Do Not Sell or Share My Personal Information Cookie Settings