Apify
Also known as: Apify Actors, Apify Store
Cloud platform for web scraping and automation with 45,000+ prebuilt Actors, proxies, and an MCP server that exposes them as tools for AI agents.
Apify is a cloud platform for web scraping, browser automation, and data extraction, built by a company of the same name based in Prague. Its core unit is the Actor, a serverless cloud program that can be triggered from the console, a REST API, the JavaScript or Python SDK, the command line, or a schedule. The Apify Store offers more than 45,000 prebuilt Actors that scrape common targets such as Google Maps, Instagram, TikTok, Amazon, and LinkedIn, so a team can ship a working scraper without writing code.
Around the Actors sits a full data infrastructure layer. The platform manages datasets, key value stores, and request queues for run output, plus a proxy network with datacenter, residential, and SERP options for sites that block automated traffic. Every run produces logs, run history, and a usage chart, so teams can debug a failing scrape and track cost without building their own monitoring. Apify also maintains Crawlee, an open source crawling library, and ships SDKs and templates in Python, JavaScript, and TypeScript for developers building custom Actors.
For agent builders, Apify positions itself as the data and tool layer rather than the agent. It runs an MCP server that exposes Store Actors as callable tools, so an assistant can invoke a scraper directly without the developer managing API keys, polling, or dataset retrieval. The company markets dedicated paths for generative AI training data and for AI agents, and counts Intercom Fin among customers pulling live web data through the platform. Apify Partners and a professional services arm also build custom scraping solutions for enterprises that prefer managed delivery.
Commercially, Apify is self serve. A permanent free plan provides $5 of monthly platform usage with no credit card, and paid plans run from $29 to $999 per month, each bundling a matching pool of prepaid usage plus pay as you go overage. Billing is usage based around compute units, proxy traffic, storage, and per Actor fees, which makes spend hard to predict for spiky or proxy heavy workloads. Apify holds SOC 2 Type II and aligns with GDPR, with single sign on reserved for the Enterprise tier.
Vendor details
Canonical URL
https://apify.com
Category
Agent infrastructure
Subcategory
Web data extraction
Company status
independent
Use cases & customers
Primary use cases
Target customers
Deployment options
Integrations
REST API, JavaScript and Python SDKs, CLI, webhooks, an integrations catalog (Zapier, Make, n8n, vector databases, LLM frameworks), and an MCP server exposing Actors as tools.
In practice
You need structured data from a site with no API and don't want to maintain scrapers. You run a prebuilt Apify Store Actor on a schedule and pull clean datasets through the REST API.
Your AI agent needs live web data mid task. You connect the Apify MCP server so the agent calls a Store Actor as a tool and gets fresh scraped results back, with no key wiring or polling.
You are blocked by anti bot defenses on a target site. You route the run through Apify's residential or SERP proxies, rotate IPs automatically, and keep the scrape alive without managing proxy infrastructure yourself.
Sources & related URLs
Related / legacy domains
Capability coverage
9.0 / 14 capabilities · 64%
| Integrations & Tool CallingPublic REST API, JavaScript and Python SDKs, CLI, an integrations catalog (Zapier, Make, n8n, vector databases, LLM frameworks), webhooks, and an MCP server exposing Actors as callable tools. | Full |
|---|---|
| Workflow OrchestrationActors can call other Actors and run on schedules with request queues and automatic retries, but orchestration is code driven rather than a visual branching workflow engine. | Partial |
| Knowledge Grounding & RAGProvides a Website Content Crawler and integrations to vector databases and LLM frameworks to feed retrieval pipelines, but is a data source rather than a built in grounding layer. | Partial |
| Human Oversight & GuardrailsNo approval checkpoints, consent gates, or autonomy controls; Apify is data infrastructure rather than an agent decisioning layer. | Unable to verify |
| Security, Identity & GovernanceSOC 2 Type II and GDPR with a public trust center; organization accounts with role based access; single sign on on the Enterprise tier. | Full |
| Observability & AuditabilityEvery run exposes logs, run history, status, and usage charts; datasets and storage are inspectable through the console and API. | Full |
| Memory & State PersistenceKey value stores, datasets, and request queues persist state across runs, though these are storage primitives rather than an agent memory layer. | Partial |
| Deployment & Data ResidencyManaged platform is SaaS only; the open source Crawlee library can be self hosted but not the Apify platform; data residency options appear limited to Enterprise. | Partial |
| Prebuilt Agents, Templates & PacksApify Store offers over 45,000 prebuilt Actors plus official Actor templates in Python, JavaScript, and TypeScript. | Full |
| Triggers & Channel CoverageRobust triggers via schedules, webhooks, REST API, and integration events; native conversational channels such as chat, email, and voice are out of scope as data infrastructure. | Partial |
| Model Flexibility & RoutingModel selection and routing are not a documented platform capability; Apify is model agnostic data infrastructure, though some individual AI Actors accept user provided model keys. | Unable to verify |
| APIs, SDKs & MCP ExtensibilityPublic REST API, JavaScript and Python SDKs, CLI, and an MCP server; developers can build, publish, and monetize custom Actors. | Full |
| Testing, Debugging & OptimizationTest runs, per run logs, and automatic retries support debugging; there is no built in output quality scoring or optimization loop. | Partial |
| Browser & Computer UseRuns headless browser automation (Playwright and Puppeteer via Crawlee) at scale with anti blocking and proxy rotation; browser based control of sites without APIs is a core capability. | Full |
Pricing
From $29/mo · free tier
compute units plus proxy GB, storage, and per Actor fees
Included quota
Each plan includes a prepaid usage pool equal to its monthly fee ($5 Free, $29 Starter, $199 Scale, $999 Business) spendable on Actor runs, proxies, storage, and Store rentals. Resource ceilings rise by tier: Actor RAM 16/64/256/512 GB, max concurrent runs 25/32/128/256, datacenter proxy IPs 5/30/200/500.
What is public
Apify publishes a complete pricing table: four self serve tiers (Free $0, Starter $29, Scale $199, Business $999) plus custom Enterprise, with a 10% annual discount. Per-unit rates are public for compute units ($0.13 to $0.20 per CU), residential proxy ($7 to $8 per GB), SERP proxy, datacenter proxy, dataset, key value store, request queue, and data transfer.
Billing mechanics
Each plan bundles a prepaid usage pool equal to its monthly fee, spendable across Actor runs, proxies, storage, and Store rentals. A compute unit is 1 GB of RAM for 1 hour. Once the pool is spent, paid plans bill overage pay as you go; the free plan blocks until the next cycle. Unused prepaid usage expires each month.
Cost watchouts
Residential proxy at $7 to $8 per GB is the most expensive line and can dominate the bill. Per Actor Store rental and per result fees stack on top of compute. Prepaid usage does not roll over, so unused budget is lost each month. Free plan hard stops with no soft overage.
Variable cost rationale
The monthly fee is only a prepaid floor. Real cost is driven by compute units ($0.13 to $0.20 per CU), residential proxy at $7 to $8 per GB, storage, data transfer, and per Actor rental or per result fees that stack on top. Spiky or proxy heavy workloads can far exceed the plan price.
Additional watchouts
The Apify Store is a marketplace: many third party Actors carry their own monthly rental or per result fees on top of platform compute, and quality and trust signals across Store Actors are uneven.
Overage / add-ons
Paid plans continue past the prepaid pool and bill overage pay as you go up to a configured limit. The free plan blocks new runs until the next cycle. Unused prepaid usage expires each month and does not roll over.
Sales call required
Mixed (some tiers require a call)
Free / trial
Free plan: $5/mo prepaid platform usage, no credit card
Lowest paid plan
Starter $29/mo ($26.10/mo billed annually) plus pay as you go usage
Commercial notes
Apify positions itself as web data infrastructure for AI agents and generative AI, with an MCP server exposing Store Actors as tools. It holds SOC 2 Type II and aligns with GDPR. Student and startup discounts of about 30% are available and nonprofits get a personalized discount. Single sign on is Enterprise only.
Key ambiguities
Total monthly cost is hard to forecast because it depends on per Actor pricing models, proxy mix, and run resources rather than the plan fee alone.
Cancellation / refund
Self serve plans can be upgraded, downgraded, or cancelled anytime in billing. Upgrades are prorated; cancellation runs to the end of the current cycle. Overage may still be charged.
Support SLA / resale
Support scales by tier: community, chat, priority chat, then a dedicated account manager on Business. SLAs with guaranteed data are Enterprise only.
Missing data
Enterprise pricing and SLAs are not public. Exact per Actor Store fees vary by Actor and are only visible on each Actor's page.
Related vendors
- AgentOps — Agent observability and reliability platform with broad model and…
- Agno — High-performance agent runtime and framework (formerly Phidata) with…
- Arcade — Authenticated tool calling platform and MCP runtime that handles…
- Arize AI — AI observability and evaluation platform that traces, evaluates, and…
- Braintrust — AI evaluation and observability platform with self-serve pricing,…
- CalypsoAI — Enterprise AI security platform that red teams, defends, and…