Introducing Server Actors: a new way to use Apify tools

Server Actors let you call Apify tools over HTTP and get results straight back, for apps, AI agents, and workflows that need live web data.

Server Actors (formerly Actor Standby mode) let you call Apify's ready-to-run tools over HTTP and get the results straight back. They run as web services that respond as soon as a request arrives, so you can build them into an application, connect them to an AI agent, or add them to any workflow that needs live data.

Apify Store has thousands of these tools, which we call Actors. They search the web, turn pages into clean text, enrich leads, monitor reviews, and more. Browse the ones that work as Server Actors, or add a server interface to your own Actor.

Two ways to use an Actor

Every Actor has an input interface: you give it an input, and it starts a run that saves its output to a dataset. That suits batch jobs and large workloads. The server interface handles one call at a time and returns the result in the response.

For example, processing 1k Facebook pages from a CSV overnight is a job for the input interface. Looking up one profile when a new lead enters your CRM is a job for the server interface.

Creators decide whether to enable the server interface. When an Actor offers both, pick the one that fits the task.

Input interface Server interface
You send An input that starts a run A request to the Actor's server URL
You get Results in the run's dataset A response, straight back
Best for Many items, crawling, anything that produces a dataset A single lookup, one page, or one object
Data Stored on Apify, where you can download it Not stored; you keep what you need

How Server Actor usage works

Server Actors support two tenancy modes: multi-tenant and single-tenant. The difference affects where your requests are processed, whose resources are used, and how you're charged.

Multi-tenant Actors run on Apify, but are hosted under the creator's account and serve requests from multiple users, like any other API. You pay per event, according to the Actor's pricing, such as per search or per review returned. The creator covers the cost of serving requests, so your calls don't use your account's memory or concurrent-run limits.

Single-tenant Actors run in your own account, and you can configure them for your own use. They're a good fit for internal tools and services your team relies on. You pay for platform usage while they run, plus any event-based charges.

Check each Actor's page and pricing details to see its tenancy model and billing.

Calling a Server Actor

If you're already familiar with APIs, using a Server Actor will feel natural. Send an HTTP request to the Actor's server URL, provide the required parameters, and get the result in the response.

For example, Web Search supports the server interface. Its server URL is https://web-search.apify.actor/, and the response depends on the parameters you send:

curl -s "https://web-search.apify.actor/?query=apify&maxResults=3&countryCode=US" \
  -H "Authorization: Bearer $APIFY_TOKEN" \
  | jq -c '.results[] | {title, url}'

{"title":"Apify: Marketplace of ready-to-run tools for AI","url":"https://apify.com/"}
{"title":"Apify","url":"https://github.com/apify"}
{"title":"Apify","url":"https://www.linkedin.com/company/apify"}

Web Search is one example of what's available. Web Fetch, for instance, returns a page as clean text, and Google Maps Reviews Scraper returns the reviews for a place. To discover more tools, browse Server Actors in Apify Store. Then open an Actor's Server tab in Apify Console to understand the expected responses and try a request.

Server tab of the Web Search Actor in Apify Console, showing the server URL and GET and POST endpoints
Web Search Server tab

Calling through MCP

Server Actors can also be used through MCP, making them available as tools to AI agents and other MCP-compatible applications.

You can connect to the Apify MCP server to discover and use Actors from Apify Store, or configure the connection to expose only the specific Actors your agent needs.

πŸ”—
Learn more in the Apify MCP documentation β†’

Calling from workflow automation platforms

Server Actors can also be invoked from workflow automation platforms such as n8n, Make, and Zapier. Use an HTTP Request step to send a request to the Actor's server URL, passing the required parameters and your Apify API token.

For example, this n8n template combines a regular HTTP request with Web Fetch. It first tries to retrieve a webpage directly and calls Web Fetch if that request fails.

Here are a couple of other examples:

  • Watch your reviews. Every morning, send a request to Google Maps Reviews Scraper for the newest reviews of your business, keep the ones under four stars, and email them to yourself.
  • Enrich a new lead. A row lands in your CRM. One call to Contact Details Scraper with the company's website returns emails, phone numbers, and social links; one to Facebook Pages Scraper adds followers and the page's own phone and email. Both are written back before the rep opens the record.

Build and publish your own Server Actor

Server Actors also give creators a way to make their tools available as callable services on Apify Store.

Building a new Actor?

To create a new Actor that supports the server interface, you can start with one of the Apify templates for JavaScript, TypeScript, or Python.

πŸ”—
To learn more about developing Server Actors, read the developer documentation β†’

Already have an Actor on Apify Store?

Your Actor may be a good fit if:

  • Many users can call it. Instead of each caller running a separate instance, multiple callers can share your infrastructure.
  • Each request returns a focused result. Profile lookups, price checks, search queries, and other tasks that return a single object are natural API use cases.
  • The work is short. When an Actor performs a small task, instance startup can make up a significant part of the total response time. Keeping instances available can help reduce that overhead.

You can enable the server interface in your Actor's settings.

Server settings in Apify Console with the Enable server interface toggle, server tenancy, requests per run, idle timeout, and memory

What to set up before you publish

  • Document your server interface. Define an OpenAPI 3 schema in webServerSchema to document how to call your Actor and what it returns. This populates the Server tab and helps AI agents call it. If your Actor exposes MCP, set webServerMcpPath to make it discoverable through Apify MCP.
  • Set up pay-per-event (PPE). Configure PPE monetization to charge callers for using your Actor.
  • Plan for resource usage. In multi-tenant mode, requests from all callers count toward your account's concurrency and memory limits. Configure your resources and scaling settings to handle expected traffic.

Already run an API or an MCP server?

Publish it on Apify Store as a Server Actor and make it available to developers looking for tools to integrate into their applications, workflows, and AI agents. Apify provides a marketplace and usage-based billing infrastructure, helping you reach new customers without building your own billing system.

Try Server Actors today

Server Actors are available in Apify Store today, and the community will keep adding more. Pick one for your next integration, or publish your own.

Apify logo
Make your first Server Actor call
Call ready-to-run Apify tools over HTTP and get the data straight back.
Browse Server Actors
On this page

Don't build it. Find it.

Thousands of ready-to-run tools for AI.

Browse Apify Store