Apify Review 2026: Pricing, Best Actors, MCP, and Real Limits

This article contains affiliate links — I may earn a commission at no extra cost to you. Disclosure

Is Apify Worth It in 2026?

Apify is worth trying when a maintained Actor already matches your job or when you need scraping, scheduling, storage, and API delivery in one managed platform. It is less attractive for a one-off scrape, a simple URL-to-Markdown request, or a workflow where unpredictable usage would be expensive.

The strongest reason to choose Apify is not that it can scrape almost anything. It is the combination of:

ready-to-run Actors for specific websites and data jobs;
managed runs, proxies, schedules, datasets, logs, and webhooks;
an API for turning a tested run into a repeatable workflow;
an MCP server that lets compatible AI clients discover and run eligible Actors.

The main weakness is the same marketplace that makes Apify useful: Actor quality, maintenance, output schemas, and pricing vary. You are choosing both the Apify platform and a particular Actor. A good platform does not rescue a poorly maintained Actor.

The decision in one table:

Choose Apify when	Choose another route when
A proven Actor already covers the target	An official API provides the required data
The workflow must run on a schedule	You only need a few manual records
You want API, storage, logs, and retries together	You only need URL-to-text or URL-to-Markdown
Your team can validate output and monitor usage	Scraping is core proprietary infrastructure you want to own
A managed pipeline is worth more than the lowest unit price	Cost must remain fixed under unpredictable volume

If the Actor works but the surrounding pipeline does not, I build web scraping automation workflows with validation, scheduling, deduplication, and delivery included.

Apify Pricing: Current Plans and the Part People Miss

Apify’s current plans range from Free to Business, but the subscription price is prepaid usage rather than a fixed all-inclusive bill. The official Apify pricing page listed these monthly plans on August 26, 2026:

Plan	Monthly price	Included prepaid usage	Compute price	Concurrent runs
Free	$0	$5	$0.20 per CU	5
Starter	$29	$29	$0.20 per CU	32
Scale	$199	$199	$0.16 per CU	128
Business	$999	$999	$0.13 per CU	256

One compute unit, or CU, represents 1 GB of memory used for one hour. Compute is only part of the bill. Residential proxies, SERP requests, storage operations, data transfer, and the selected Actor’s own pricing can also consume the prepaid amount.

Three details matter more than the headline plan price:

Prepaid usage is spend, not a separate allowance for every service. Actor runs and platform resources draw from the same plan balance.
Unused prepaid usage expires at the end of the billing cycle. It does not roll over.
Overage depends on the account limit. Apify suspends platform services when the configured usage limit is reached. A paid account can raise that limit and enable pay-as-you-go; usage beyond the plan is then billed as overage.

Actor pricing is separate from plan pricing

Store Actors generally use one of two current models:

Pay per event: the developer charges for defined events such as starting a run, producing a result, or processing an item. Most include platform usage, but some add it separately.
Pay per usage: there is no separate developer fee; you pay for the compute, proxy, storage, and transfer resources consumed by the run.

Do not rely on an old price quoted in an Actor README. The current plan page and the Actor’s Pricing tab are the authorities. Some Actor descriptions captured during this review contained CU examples that did not match the live platform table.

The useful cost question is not “What does Apify cost?” It is “What does this Actor cost per accepted record after retries, duplicates, proxy traffic, and review?” My web scraping pricing guide provides a broader comparison framework.

Five Apify Actors Worth Starting With

Start with an Actor that already has meaningful current usage and matches a specific business decision. Selection is based on job relevance, current Store adoption, recent maintenance, rating volume, and a pricing contract visible before running, not affiliate payout. A large user counter shows adoption, not data quality.

The Store numbers below are live snapshots captured on August 26, 2026. They will change.

Actor	Best for	Total users	Monthly active	Rating	Listed price
Google Maps Scraper	Local-business and B2B research	577K	35K	4.7 (1,782)	From $1.50 / 1,000 places
Instagram Scraper	Brand, creator, and social research	375K	41K	4.7 (573)	From $1.50 / 1,000 results
Google Search Results Scraper	SEO, ads, and AI-search monitoring	171K	15K	4.6 (184)	From $1.80 / 1,000 result pages
TikTok Scraper	Short-form trends and creator research	244K	20K	4.8 (363)	From $1.70 / 1,000 results
Website Content Crawler	Documentation, RAG, and content corpora	150K	9.8K	4.6 (223)	Pay per usage

These are starting points, not unconditional winners. Run the same small canary you would use to evaluate any third-party data provider.

1. Google Maps Scraper for local-business research

The Google Maps Scraper is the clearest commercial starting point in the reviewed Store data. It extracts places, categories, ratings, reviews, coordinates, contact fields, opening hours, and other business attributes. Compass is the listed developer; the Store marks the Actor as maintained by Apify.

Use it when the decision begins with a defined market such as “dentists in Almaty” or “SaaS companies in Berlin,” not when the plan is to collect an unlimited list and decide what to do later.

A responsible lead workflow still needs:

a specific geography and business category;
required fields and a rejection rule;
deduplication by place ID or another stable identifier;
verification before outreach;
a lawful basis for any personal-data processing.

See the Google Maps lead-extraction guide for the workflow design beyond the Actor run.

2. Instagram Scraper for brand and creator research

The Instagram Scraper covers public profiles, posts, reels, comments, hashtags, mentions, places, and related metadata. Its 41K monthly active users were the highest among the five shortlisted Actors in this snapshot.

It is useful for bounded tasks such as:

comparing a fixed set of competitor accounts;
collecting posts under a defined campaign hashtag;
reviewing creator output before a partnership decision;
tracking public brand mentions over a stated time window.

It is not a substitute for an Instagram account database or permission to repurpose personal data. Save source URLs and retrieval times so findings can be checked later.

3. Google Search Results Scraper for SEO and AI-search monitoring

The Google Search Results Scraper extracts organic results, paid results, People Also Ask, related queries, product elements, and newer AI-search surfaces. Its Store page also lists optional Google AI Mode, Perplexity, ChatGPT search, Copilot, and Gemini modes for cross-platform comparison. Treat these as separate options with their own pricing conditions, not as proof that one standard SERP request covers every engine.

This is relevant for:

checking a defined keyword and country set;
monitoring paid and organic competitors;
capturing AI-answer citations over time;
feeding a rank or narrative-change report.

Do not call a single scrape a ranking trend. Keep the language, country, device assumptions, timestamp, and exact query beside each result. For a smaller owned implementation, see the Google SERP checker.

4. TikTok Scraper for short-form trend research

The TikTok Scraper extracts public video, profile, hashtag, engagement, and music-related data. Clockworks is the listed developer; the Store marks the Actor as maintained by Apify.

The Actor is useful for discovering patterns across a controlled set of creators or hashtags. It is weaker as a source of universal market truth. Platform search and recommendation systems change, and different collection methods can return different datasets for the same nominal query.

For a trend study, preserve the query, collection time, result limit, sort mode, and missing records. Without that context, a spreadsheet of TikTok rows looks more authoritative than it is.

5. Website Content Crawler for RAG and documentation

The Website Content Crawler crawls a site and converts pages into cleaned text or Markdown for LLM, vector-database, and RAG workflows. It supports raw HTTP and browser-based crawling, file downloads, URL controls, and integrations with common AI tooling.

This Actor is a good fit for documentation, help centers, blogs, and knowledge bases with multiple related URLs. Its own documentation recommends a lighter single-page tool when you only need one URL. That is an important boundary: do not pay for a full crawl when the task is one fetch.

Before embedding the output, remove navigation and duplicate templates, preserve canonical URLs, and reject pages with missing titles or empty bodies. The RAG Web Browser guide covers the narrower search-and-fetch route.

When the general Web Scraper makes sense

The Web Scraper had 124K total users, a 4.7 rating, and 1.3K monthly active users in the same snapshot. It renders pages in a browser and runs custom JavaScript extraction logic.

Use it when no specialized Actor fits and someone on the team can maintain selectors, page functions, and debugging. A specialized Actor is usually a faster first test; the general Web Scraper is the controlled fallback, not the default recommendation for a nontechnical buyer.

How to Choose Between Similar Actors

Choose between similar Actors by output fit, maintenance, current usage, pricing clarity, and a controlled test run. Searching the Store by name often returns several Actors that claim to scrape the same target. Price alone is a poor tie-breaker.

Use this order:

Match the output to the decision. List the fields you require before opening the Store.
Prefer a maintained Actor with meaningful current usage. Total users show history; monthly active users are a better freshness signal.
Read the Pricing tab. Identify every paid event and whether platform usage is included.
Inspect the output schema and a real dataset. Check types, nulls, source URLs, timestamps, and identifiers.
Read recent issues and the changelog. A high rating does not reveal a current breaking change.
Run the same canary across finalists. Compare accepted records and total billed usage, not raw row count.

The platform’s marketplace counters are useful filters, not proof. A smaller specialized Actor can beat a popular general Actor on one target. A heavily used Actor can also inherit the limits of the website it collects from.

A practical canary

Use an input small enough to inspect manually:

Record	What to save
Input	Exact JSON, query, limits, locale, and target URLs
Run	Actor ID, run ID, version, status, and elapsed time
Output	Returned rows, accepted rows, duplicates, and missing fields
Evidence	Source URL and retrieval time for every important record
Billing	Actor events, compute, proxy, storage, transfer, and retries
Decision	Pass/fail rule and the reason for rejection

One verified canary is more informative than a page of generic feature claims.

Apify MCP: Useful, but Not Magic

Apify MCP gives AI clients a controlled way to discover and run eligible Actors, but it does not remove cost or data-quality risk. The official Apify MCP Server also lets clients inspect runs, retrieve datasets, and search Apify documentation.

The recommended hosted connection uses OAuth. Token authentication is also available, and the tool set can be restricted to specific capabilities or Actors. A narrow configuration such as Instagram plus Google Search is safer and easier for an agent to use than exposing the entire Store.

Important limits from the official documentation:

full-permission Actors are excluded because an LLM must not approve that access on your behalf;
rental Actors are excluded from MCP execution;
Actor pricing and platform usage still apply;
the server allows up to 30 requests per second per user;
MCP changes the interface, not the reliability of the target or the correctness of the data.

Apify also has “MCP connectors,” which solve the opposite problem: they let an Actor call an external MCP service while Apify injects approved credentials server-side. Do not confuse connectors with the public Apify MCP server.

MCP makes an Actor callable by an AI agent. It does not make the Actor trustworthy, free, or autonomous enough to skip validation.

Real Advantages and Real Friction

Apify saves infrastructure work when a suitable Actor exists, while shifting quality, schema, and cost control onto the team running it.

Where Apify is strong

A specialized Actor can avoid building and operating a target-specific integration from scratch.
Runs, datasets, logs, schedules, webhooks, proxies, and API access live in one system.
JavaScript and Python teams can build custom Actors without operating the runtime infrastructure.
The Store creates options for difficult sources that have weak or restricted official APIs.
MCP makes the same Actors accessible to agent workflows without inventing a separate tool protocol.

Where teams get surprised

Compute, proxy traffic, retries, storage, and Actor events make forecasting harder than reading the plan price.
Community Actors vary in maintenance, documentation, deduplication, and schema stability.
A run can finish successfully while returning incomplete or unusable records.
Multi-Actor chains need your own acceptance rules, observability, and failure ownership.
The interface is approachable for a first run, but production configuration still becomes technical.

This split appears in independent user evidence. Capterra showed Apify at 4.8 from 559 reviews on August 20, 2026, while verified reviews still described third-party schema drift, duplicate charges caused by Actor behavior, technical onboarding, and the need to monitor proxy and compute usage. A competing ScrapingBee review independently praised the platform’s flexibility while warning about uneven Actor quality and hard-to-predict compute spend.

Those sources have their own commercial incentives. They are used here to identify recurring objections, not as the authority for Apify’s current prices or product contract.

Apify Alternatives by Operating Model

The right Apify alternative depends on which part of the platform you do not need. A direct API, visual tool, official source API, or self-hosted crawler solves a different operating problem.

Need	Better first option	Why
Stable data already exposed by the source	Official API	Clearer contract and less breakage
One URL or a simple crawl returned as Markdown	Direct scraping API	Smaller surface and simpler billing
A nontechnical one-off workflow	Visual extraction tool or manual research	Less setup and maintenance
A recurring target-specific job	Maintained Apify Actor	Fast path from test to scheduled run
Proprietary extraction logic on managed infrastructure	Custom Apify Actor	Control without running the platform
Scraping is core infrastructure at sustained scale	Self-hosted Crawlee, Scrapy, or browser stack	Maximum control and portability
Enterprise proxy and dataset operations	Specialized data/proxy provider	Procurement, coverage, and service-level options

Apify is overkill when the job is “fetch this page and return clean text.” It becomes more compelling when the job is “run this source-specific extraction every morning, retain the dataset, retry failures, and send accepted rows into another system.”

Buying Recommendation

Start on Free and upgrade only after one bounded run proves output quality and measured cost. Choose one of the five Actors above because it matches a real job, not because it appears high in Store search. Set a small result limit, run it once, inspect every returned field, and reconcile the usage ledger.

Move to Starter only when one of these is true:

the $5 Free allowance blocks a validated recurring workflow;
you need more than five concurrent runs;
chat support matters or a controlled pay-as-you-go allowance is operationally useful;
the Actor’s measured monthly cost fits inside a defined budget.

Do not upgrade because a larger plan looks cheaper per CU. Scale and Business make sense after measured volume, not before it.

If the canary fails on output quality, maintenance, or cost, test one competing Actor with the same input. If both fail, change the operating model rather than endlessly swapping marketplace listings.

FAQ

Is Apify free?

Yes. The current Free plan costs $0 and includes $5 of monthly prepaid usage. It has five concurrent runs and a $0.20 CU rate. When the free usage is exhausted, access is blocked until the next cycle; there is no paid overage on Free.

Which Apify Actor should a marketer try first?

Choose by the decision you need to make. Google Maps Scraper fits local-business research, Instagram or TikTok Scraper fits public social analysis, Google Search Results Scraper fits SEO and AI-search monitoring, and Website Content Crawler fits documentation or RAG corpora.

Are popular Actors automatically reliable?

No. Total and monthly users show adoption, not field accuracy. Check recent issues, output schema, maintenance, pricing, and a manually verified canary.

Does Apify MCP include every Actor?

No. The official MCP server excludes full-permission and rental Actors. You can also restrict the available tool set to specific Actors.

Can Apify costs exceed the subscription price?

Yes. Paid accounts can raise their platform usage limit and enable pay-as-you-go; usage beyond the subscription is then charged as overage. Keep the limit bounded and reconcile the first run before scheduling.

Is Apify better than Firecrawl or ScrapingBee?

Apify is broader: it combines a marketplace, cloud runtime, storage, schedules, and custom code. A direct scraping API can be simpler for URL-to-content jobs. Compare the operating model before comparing per-page prices.

Is web scraping legal?

There is no universal answer. It depends on the source, access method, data, jurisdiction, terms, and intended use. An Apify Actor provides infrastructure; it does not grant permission to collect or use data.

Sources and Verification

Current product and pricing claims were checked against:

Apify pricing
Apify billing and platform limits
Actors in Store pricing models
Pay-per-event documentation
Apify MCP documentation
the five linked Actor Store pages and their Pricing tabs
Capterra’s Apify review page
ScrapingBee’s 2026 tool comparison

Last verified: August 26, 2026. Store counters and prices are snapshots; check the linked pages before purchasing or scaling a run.