Apify API review
Best for a scraper somebody already built
POST https://api.apify.com/v2/actors/compass~crawler-google-places/run-sync-get-dataset-items Authorization: Bearer $APIFY_TOKEN # Google Maps Scraper: from $1.50 per 1,000 places # billed per event by the Actor, not per compute unit # sync calls stop at 300 seconds; use /runs beyond that
The Apify API is not a scraper. It is the control plane for Apify Store, a marketplace of thousands of Actors that other people wrote and maintain, from a Google Maps scraper to a whole-site crawler. You send one token, start an Actor with JSON input, and read the results out of a dataset, and the same calls work for an Actor you wrote yourself.
That shape decides the bill. A plan is prepaid platform usage in dollars, not a quota of scrapes: $5 a month free, then $19, $199 or $999, and whatever you do not spend expires at the end of the cycle. What one scrape costs is set by the Actor, and most Store Actors charge per event, so Google Maps Scraper is from $1.50 per 1,000 places whatever it takes to fetch them.
The useful consequence is that the Apify API cost of a job is readable before you run it, on the Actor's own page. The trap is that the same API also runs pay-per-usage Actors, where you pay compute units instead, and those cost whatever the run happens to take.
The rest of the category is in the Sourcing APIs listing.
What one scrape costs, and who sets the price
Every Actor page states one of two pricing models. Pay per event is a fixed price per result or action, set by the Actor's developer, with platform usage usually included. Pay per usage charges nothing on top and bills the compute units, proxies and storage the run consumed.
| What you run | Pricing model | What Apify publishes | Where it comes from |
|---|---|---|---|
| Google Maps Scraper | Pay per event | from $1.50 per 1,000 places | The Actor's own page, lower on higher plans |
| Its add-ons | Pay per event | $0.20 per 100 place details | Free-plan rate; $0.15 on Silver |
| Website Content Crawler | Pay per usage | $0.20 per 1,000 pages, raw HTTP | Apify's tests, on the Actor's page |
| The same, in a browser | Pay per usage | $0.50 to $5 per 1,000 pages | Depends on the site, per the same tests |
| Your own Actor | Compute units | $0.20 to $0.13 a CU | 1 CU is 1 GB of RAM for 1 hour |
| Google results through Apify Proxy | Proxy | $2.50 to $1.70 per 1,000 SERPs | By plan, from the pricing page |
The two models answer different questions. Pay per event tells you the price of a result, which is what a pipeline wants to forecast. Pay per usage tells you the price of a minute of a machine, so a slow or heavily defended site costs more for the same row, and Apify's own FAQ says retries on failed requests add to it.
The pricing page is the rate card: $0.20 a compute unit on Free and Starter, $0.16 on Scale and $0.13 on Business. The Website Content Crawler page works out its per-1,000-page estimate at a baseline of $0.25 a unit, which no current plan charges, so on Starter the same run should cost about a fifth less. Run a test with a small input and take the real figure from the usage chart in Apify Console, which is the advice Apify gives itself.
Apify API pricing, plan by plan
Apify API pricing is a monthly fee that comes back to you as the same amount of prepaid platform usage, with overage billed on paid plans. Annual billing saves 10%, and the annual monthly rate is shown beside each price.
Free
- No credit card
- Compute at $0.20 a CU, 16 GB of RAM
- 5 concurrent runs
- Blocked, not billed, when the $5 runs out
Starter
- $19 of prepaid usage a month
- Compute at $0.20 a CU, 64 GB of RAM
- 32 concurrent runs
- Bronze discount on Store Actors
Scale
- $199 of prepaid usage a month
- Compute at $0.16 a CU, 256 GB of RAM
- 128 concurrent runs
- Silver discount on Store Actors
Business
- $999 of prepaid usage a month
- Compute at $0.13 a CU, 512 GB of RAM
- 256 concurrent runs
- Gold discount and an account manager
None of these tiers is a number of results, so this directory does not model Apify by volume. The honest conversion runs through an Actor: $19 of usage buys about 12,700 Google Maps places at the listed $1.50 per 1,000, before any add-on and before the Bronze discount. A different Actor gives a different number, which is the whole point of the model.
Apify says it plainly: unused prepaid usage is not rolled over and disappears at the end of the billing cycle. A $199 plan that spends $60 in a month has paid $139 for nothing, so size on the month you actually run rather than the month you hope to.
Per resource, per account, and generous
Apify API rate limits come in two layers, and neither is likely to be the thing that slows a GTM pipeline. The global limit is 250,000 requests a minute, counted per user. The per-resource limit is 60 requests a second against a single Actor, run, dataset or store, and a handful of endpoints are allowed more.
| Endpoints | Limit per resource |
|---|---|
| Run an Actor, run a task, metamorph a run | 400 a second |
| Request queue record operations and locks | 350 a second |
| Key-value store record operations, list keys | 200 a second |
| Key-value store download as a ZIP | 100 a second |
| Everything else, by default | 60 a second |
| Request queue batch operations, list and lock head | 35 a second |
| The whole account | 250,000 a minute |
What actually caps throughput
Runs, not requests.
- Concurrent runs: 5 on Free, 32 on Starter, 128 on Scale, 256 on Business.
- Memory: 16 GB on Free up to 512 GB on Business, shared by everything running.
- Extra concurrent runs are $5 each and extra RAM is $1 a GB.
- A synchronous run returns at 300 seconds; longer work goes through the async flow.
When a call is refused
A 429, with the limit named in the body.
- Each endpoint returns its limit in the X-RateLimit-Limit header.
- Exceeding it returns 429 and rate-limit-exceeded.
- Apify asks for exponential backoff: wait, retry, double the wait.
- The per-resource limit is per Actor or dataset, so fan out across resources, not calls.
Apify rate limits are the least of the ceilings here. The constraint a pipeline meets first is concurrent runs on the plan, and after that the prepaid usage running out, which on Free stops the account until the next month.
How to create an Apify API key
Asking how to get Apify API token access and how to get Apify API key access is asking one question. Sign in to Apify Console and open the API and Integrations section of Settings, where the token is waiting, with no card needed on the free plan. The Apify API key and the Apify API token are the same credential under two names.
The token, for code
One credential for the REST API and both clients.
- The Apify API token for authentication goes in Authorization: Bearer <token>.
- A ?token= query parameter also works, and Apify calls it less secure.
- Use apify-client to call Actors; the apify package is for building them.
- Bill caps travel with the call: maxTotalChargeUsd and maxItems.
The MCP server, for an assistant
Hosted at mcp.apify.com, over streamable HTTP.
- Using the Apify API for Claude or Cursor: add https://mcp.apify.com as a server.
- The first connection signs in with OAuth, so nothing is pasted.
- Or pass the Apify API token for Claude as a Bearer header instead.
- apify mcp install cursor writes the configuration for you.
There is a third route, which BuiltWith and Browserbase also offer in their own form: an agent with a crypto wallet can buy a spend-capped token at agi.apify.com through x402 or MPP, with no account at all. For a human team the token from Console is the answer, and the bill caps on each call are the habit worth forming.
A first Apify API call
These are the calls from Apify's own agent quickstart. The synchronous endpoint starts the Actor and returns its dataset items in one response, which suits anything under 300 seconds; the asynchronous flow starts a run, polls it and then reads the dataset.
# short jobs: run and read in one call POST https://api.apify.com/v2/actors/compass~crawler-google-places/run-sync-get-dataset-items # longer jobs: start, poll, then read POST https://api.apify.com/v2/actors/compass~crawler-google-places/runs # poll GET /v2/actor-runs/{runId} until SUCCEEDED, FAILED, TIMED-OUT or ABORTED # then GET /v2/datasets/{datasetId}/items, add ?format=csv for CSV
Where the Apify API documentation lives
The Apify API documentation is at docs.apify.com/api/v2, and every page has a Markdown twin: append .md and you get the source. The Apify API docs for agents live at apify.com/agents.md, which is the page to hand an assistant.
| Surface | Where | What it answers |
|---|---|---|
| API reference | docs.apify.com/api/v2 | Every endpoint, rate limits, errors and pagination. |
| Agent quickstart | apify.com/agents.md | MCP, the clients, the HTTP flow and the cost rule. |
| Pricing as Markdown | apify.com/pricing.md | Plans, compute and proxy rates in one table. |
| MCP server | docs.apify.com/platform/integrations/mcp | OAuth or token, per-client setup, tool selection. |
| Clients | apify-client for JavaScript and Python | The supported way to call Actors from code. |
An Apify API integration needs two of those: a token from the API and Integrations page in Apify Console, scoped if the job should only run certain Actors, and the apify-client library. Read the Actor's own page as part of the documentation. It carries the input schema, the output fields, the pricing model and the example costs, and for a Store Actor it is the only place the price of a result is written down.
The MCP page lists Apify's MCP access alongside the others in this directory.
Apify's own walkthrough of the API
Apify publishes this on its own channel, so it is Apify's account of Apify, not ours. It is here because a rate card cannot show you the thing moving, and a few minutes of the product is faster than any paragraph on this page at telling you whether the shape of it fits yours.
YouTube · Apify How to programmatically retrieve data with the Apify API 3:31 · Published August 2023Three and a half minutes, by Apify's own description, on getting a run's results out of Apify with cURL and with the JavaScript and Python API clients. It dates from 2023, so check any URL it shows against the current API reference, and no price on this page comes from it.
Published 8 August 2023 on Apify, the channel Apify's own pages link to, and read on 23 October 2026. Nothing in it is a source for a price or a limit on this page: every figure in the tables above comes from an Apify page.
Strengths and limits
Apify is strongest when the scraper you need already exists, and weakest when you need to know the cost of a run that nobody has priced.
What it does well
- Thousands of maintained Actors behind one token, so most sources are a configuration rather than a build.
- Pay-per-event Actors publish the price of a result on their own page.
- Bill caps per call: maxTotalChargeUsd and maxItems.
- A hosted MCP server with OAuth, and Markdown twins of the docs and pricing.
- Generous rate limits: 60 a second per resource, 400 to start runs.
- A free plan with $5 a month and no card, which stops rather than bills.
Where it falls short
- Plans are dollars of usage, not results, so the cost of a job depends on the Actor.
- Pay-per-usage Actors cost whatever the run takes, including retries.
- Unused prepaid usage expires every month.
- An Actor is somebody else's code; when the target site changes, you wait for its developer.
- Cost estimates on Actor pages can lag the rate card: the Website Content Crawler still assumes $0.25 a compute unit.
Taken together, Apify is a marketplace first and an API second. The API is excellent at starting, capping and reading runs; whether a particular scrape is cheap, reliable or maintained is a property of the Actor, and the Actor page is where that gets decided.
Who should actually buy this
Buy Apify when the source you want already has a maintained Actor. Google Maps listings, a directory, a marketplace, a job board: for most of them somebody has built and priced the scraper, and the Apify API turns it into one call with a known price per result.
Start on the free plan, with a small input and maxTotalChargeUsd set, and read the usage chart afterwards. That one run answers what the pricing page cannot, which is what your particular job costs.
Look elsewhere when the job is reading arbitrary company websites as clean text for a model, where a single-purpose scrape API is simpler to price, or when the source has no Actor and no budget for building one.
Before you build against it
Four things to settle first, each one read off Apify's own pages. The ticks are remembered in this browser, so you can come back to a half-finished integration.
Where the data comes from, and what you may do with it
The question that arrives late and expensively is who is allowed to hold the data. These are the fields it returns, the coverage it claims and the compliance it publishes.
| Provider | Fields returned | Coverage | Compliance | Redistribution |
|---|---|---|---|---|
| Apify | Firmographics, Web pages | Global | SOC 2 GDPR | Not stated |
Apify states no redistribution position on its pricing or docs pages. Internally that never matters. Put the rows in something your customers can see and it becomes the question your legal team asks first.
Two practical notes. A compliance badge describes the vendor rather than your use of it, so a GDPR DPA does not give you a lawful basis for the outreach you send. And the vendor runs its own opt-out, which means a person who asks you to delete their record has to be removed from your copy as well as from theirs.
If Apify is not the answer
Four providers in the Sourcing APIs listing that solve the same problem differently. Each one is a trade, and the trade is named.
Bright Data
$1.50/1K requestsPre-built scrapers priced by the vendor itself: $1.50 per 1,000 records on success, 5,000 free a month, and unlimited concurrency.
Firecrawl
$16/mo annualOne credit a page, rendered, returned as Markdown. The simpler price when the job is reading company websites for a model.
ScrapingBee
$19/moThe cheapest rendered page in this group: 5 credits for a default rendered fetch at $0.000099 a credit on Startup, about $0.0005.
PhantomBuster
$56/mo annualIf the Actors you ran were social and LinkedIn automations, this runs the same jobs through your own accounts, metered in execution hours.
The full set, organized by the reason people leave, is on the Apify alternatives page.
Apify against the obvious rival
Same rate cards, read side by side, with the arithmetic done on a worked volume rather than left to you: Apify vs Firecrawl, Apify vs Bright Data, Apify vs PhantomBuster and Apify vs Browse AI.
The same again, against the rest of the field: Apify vs Browserbase, Apify vs Oxylabs, Apify vs ScraperAPI and Apify vs ScrapingBee.
Every pairing in the directory is on the compare hub.
Apify vs Firecrawl
CompareA marketplace that runs other people's crawlers against one endpoint built to turn a page into Markdown. Raw HTTP is cheaper on Apify; a rendered page is close.
Apify vs Bright Data
CompareTwo platforms selling scrapers, an unblocker and proxies, and pricing the unblocker identically on the entry tier. Search is where they part.
Apify vs PhantomBuster
CompareBoth bill for running code rather than for rows. PhantomBuster sells execution hours through your own accounts; Apify sells prepaid usage and lets each Actor price its results.
Calling these APIs one by one? Run them from one Clay table.
Clay runs a waterfall across 150+ data providers from one table, without a separate contract with each. A lookup that finds nothing costs no data credit, and the free plan includes 500 actions a month.
Frontal is a Clay Elite Studio Partner.
Nothing worth quoting yet
We searched Reddit for threads about the Apify API and have found none worth quoting yet. When one turns up, it goes here, and nothing on this page leans on an unquoted paraphrase.
Apify API, answered
Is the Apify API free?
Free plan, $5 of platform usage a month, no card. Blocked rather than billed when it runs out. Paid use starts at $19/mo.
How much does the Apify API cost?
Apify starts at $19/mo. Its unit rate is $0.20 per compute unit. It bills pay as you go or by subscription, with a free tier. Read from Apify's own pricing page on 24 October 2026.
What are the Apify API rate limits?
60 requests a second per resource, 400 to run an Actor, 250,000 a minute per account.
How do I get an Apify API key?
Asking how to get Apify API token access and how to get Apify API key access is asking one question. Requests authenticate with bearer API token, or OAuth through the hosted MCP server. The current steps are in Apify's own documentation, which is authoritative on anything that has changed since 24 October 2026.
Does Apify have an MCP server?
Yes. Hosted at mcp.apify.com, OAuth or token. It is listed with the others on the MCP page.