Skip to content

Apify

Cloud platform for running scrapers and browser automations, with a store of ready-made 'Actors' for hard-to-scrape sites.

Reviewed by Andrew Miller · Updated

Visit Apify Free plan

Our rating

4.1/5

Apify turns scraping big, defended platforms into configuration rather than engineering, at the price of a bill nobody can predict before the job runs.

What it is

Apify runs scraping and automation programs, which it calls Actors, on its own cloud. An Actor takes JSON input, runs on Apify's infrastructure and writes structured output to a dataset or key-value store. The platform adds scheduling, proxies, storage and an API around them.

Apify Store lists thousands of ready-made Actors for sites such as Google Maps, Instagram and online shops. Most are built, priced and maintained by independent developers rather than by Apify. You can also write your own, usually with Crawlee, Apify's open-source scraping library for Node.js and Python. A hosted MCP server at mcp.apify.com lets AI agents find and run Actors.

Plans are Free ($5 of usage a month, no card), Starter ($19), Scale ($199) and Business ($999), or about 10% less billed annually. Each plan's fee is prepaid usage, billed in compute units (1 GB of RAM for one hour) at $0.13 to $0.20 depending on plan, with residential proxies at $7 to $8 per GB and many Store Actors charged on top.

Our review

Apify's advantage is that someone has usually already built the scraper. For Google Maps listings, social profiles or product pages, a Store Actor turns a hard engineering job into configuration, and one reviewer, an Apify partner, describes setting up a ready-made Actor in about 20 minutes. Crawlee, the open-source library underneath many Actors, has about 26,200 GitHub stars and released v3.18.2 on 29 September 2026, so the custom route is well supported too.

The weakness is cost and consistency. Your plan fee buys prepaid usage, then compute at $0.13 to $0.20 per unit, residential proxy traffic at $7 to $8 per GB and per-Actor charges stack on top. Unused credit expires at the end of each billing cycle, and paid plans bill overage on the next invoice up to a usage limit you set, so that limit is the only real budget control. Store Actors are third-party code: well-maintained ones are good, but a community Actor can break when its target site changes and stay broken until its author fixes it.

For engineers, the platform limits are generous and the storage rules are the trap. The API allows 250,000 requests a minute per user and 60 a second per resource, and a Business account can run 256 jobs at once with 512 GB of combined memory. Run results land in unnamed datasets by default, which are deleted under the plan's retention rules; on Free only the 10 most recent runs are kept, for four months. A pipeline that forgets to name its dataset or export results can lose them.

Apify holds a SOC 2 Type II report covering security, availability and confidentiality, available under NDA from its trust centre, and uses EU standard contractual clauses for data leaving the EEA. That covers the platform, not the legality of what you scrape; terms of service and personal-data rules for the target site remain the buyer's problem. For retrieval pipelines that only need arbitrary pages as Markdown, Jina Reader or Firecrawl are simpler tools with fewer moving parts than Apify's RAG Web Browser Actor.

Where it earns its place

  • A growth analyst at a 30-person B2B agency

    Pull Google Maps listings for every dentist in five US cities each month

    Run a maintained Store Actor on a schedule and export the dataset as CSV or JSON. Starter at $19 a month often covers small pulls, but per-result Actor fees and proxy traffic decide the real cost, so set a usage limit first.

  • A data engineer feeding a pricing model from 40 retail sites

    Run custom scrapers daily with proxies and retries

    Write Crawlee scrapers, deploy them as Actors and trigger runs via the API or webhooks. At 1,000 compute units and 50 GB of residential traffic a month on Scale, usage comes to about $535, so roughly $336 is billed on top of the $199 fee.

  • An AI engineer building a research agent

    Let an agent fetch structured data from social and search platforms

    Connect the hosted MCP server at mcp.apify.com so the agent can search the Store and run Actors with your token. Every run spends credit, so a looping agent needs a hard usage limit on the account.

Before you commit

Setup
Free account with $5 of monthly usage and no card; a Store Actor runs from the console in minutes. Custom Actors use the Apify CLI and Crawlee templates for Node.js or Python.
Integrations
REST API, JavaScript and Python API clients, webhooks, schedules and the hosted MCP server at mcp.apify.com; Store Actors cover most third-party platforms directly.
Data and privacy
SOC 2 Type II (security, availability, confidentiality) with the report under NDA; GDPR with standard contractual clauses. Named storages are kept indefinitely; unnamed ones follow plan retention, with Free keeping only the 10 most recent runs for four months.
Cost at scale
Scale at $199 a month covers about 1,240 compute units at $0.16; a workload using 1,000 units plus 50 GB of residential proxy costs about $535. Business at $999 lowers compute to $0.13 and proxy to $7 per GB.
Limits
250,000 API requests a minute per user and 60 a second per resource; concurrency of 5, 32, 128 and 256 runs from Free to Business, with 16 GB to 512 GB of combined memory.

What's good

  • Ready-made scrapers for many large platforms
  • Scheduling, storage, proxies and API in one place
  • Open-source Crawlee for custom scrapers
  • Hosted MCP server for agents
  • SOC 2 Type II report available under NDA

What's not

  • Compute, proxy and Actor fees stack, so bills are hard to forecast
  • Unused prepaid credit expires each cycle
  • Third-party Actor quality and upkeep vary
  • Unnamed run datasets are deleted under retention rules
  • Large jump from $199 to $999 plans

Best for

Teams that need structured data from specific large websites and would rather pay than maintain scrapers.

Alternatives

Apify: common questions

Is Apify worth it?
Apify turns scraping big, defended platforms into configuration rather than engineering, at the price of a bill nobody can predict before the job runs. We rate it 4.1 out of 5.
How much does Apify cost?
Free plan with $5 usage/month; Starter $19, Scale $199, Business $999 per month as prepaid usage (about 10% less yearly); compute $0.13–$0.20 per unit; residential proxy $7–$8/GB; Store Actors extra.
What are the downsides of Apify?
Compute, proxy and Actor fees stack, so bills are hard to forecast; unused prepaid credit expires each cycle; third-party Actor quality and upkeep vary; unnamed run datasets are deleted under retention rules; large jump from $199 to $999 plans.
Who is Apify best for?
Teams that need structured data from specific large websites and would rather pay than maintain scrapers.
How much does Apify cost for a team or in production?
Scale at $199 a month covers about 1,240 compute units at $0.16; a workload using 1,000 units plus 50 GB of residential proxy costs about $535. Business at $999 lowers compute to $0.13 and proxy to $7 per GB.
Is Apify safe for company data?
SOC 2 Type II (security, availability, confidentiality) with the report under NDA; GDPR with standard contractual clauses. Named storages are kept indefinitely; unnamed ones follow plan retention, with Free keeping only the 10 most recent runs for four months.
What are the best Apify alternatives?
The closest alternatives we have reviewed are Jina Reader.

Answers checked Oct 22, 2026.

Make Apify? Show this rating on your siteApify rated 4.1 out of 5 on ReviewRamble