Our scrapers now answer like APIs: 14 Apify Actors you can call with one HTTPS request
Ask for one LinkedIn job, one Zillow home or today's Treasury curve, and the JSON comes back in the same HTTP response. In our test calls on November 18, 2026 UTC, answers took 3.8 to 11.0 seconds for 12 of the 14. Googl
Ask for one LinkedIn job, one Zillow home or today's Treasury curve, and the JSON comes back in the same HTTP response. In our test calls on November 18, 2026 UTC, answers took 3.8 to 11.0 seconds for 12 of the 14. Google Search Results and Reddit open a real browser and took 48.9 and 21.8 seconds. Your AI agent can call these Actors in the middle of a task and n8n, Make and Zapier can use them in a single HTTP step. Your app can treat them like any other REST API.
Apify calls this a Server Actor (it was called Standby before). You don't start a run, wait for it to finish and then read a dataset. You send a request to the Actor's own address and read the answer.
What changes for you
-
Answers in seconds. The result is in the response body, under
data. You don't need to poll, and you don't need a second call to fetch results. -
Works anywhere you can send HTTPS. That covers curl, Python
requests, a function-calling tool in your agent, the HTTP Request node in n8n, the HTTP module in Make and the webhook step in Zapier. All of them work the same way: a URL, query parameters and your Apify token in theAuthorizationheader. -
A schema your tools can read. Every endpoint serves
GET /openapi.json, so an agent framework or API client can import the request and response format directly. You don't have to describe it by hand. - The same Actor still runs in bulk. The classic run with an input form, datasets and schedules works as before.
The 14 Actors you can call today
Each row is one live endpoint at https://automatic--<actor>.apify.actor/. The response time comes from one test request per Actor, made by our deploy script right after each server went live on November 18, 2026 UTC. Your times will vary with the source site and with whether the server is already running.
| Actor | You send | You get back | Test response | How it's billed |
|---|---|---|---|---|
| LinkedIn Jobs | jobId |
One job's full description, company, location, posting date and applicant count | 5.3 s | Per request: $0.001 |
| OCC Mexico Jobs |
q (keyword) or a company board url, plus limit
|
Matching Mexican job offers in OCC's displayed order | 5.9 s | Per request: $0.001 |
| Instagram Post |
shortcode or post link |
Caption, hashtags, engagement, timestamp and media URLs | 4.8 s | Per request: $0.003 |
| LinkedIn Company Posts | Company url, limit
|
The company's newest public posts with reactions and media | 4.9 s | Per request: $0.0036 |
| Zillow | zpid |
One home's full details: price, beds and baths, agent, tax, HOA, price history and photos | 5.3 s | Per request: $0.0009 |
| US Treasury Yield Curve |
date, or latest=true
|
All 13 Treasury yields and the 2s10s spread for one trading day | 3.8 s | Per row: $0.002 |
| YouTube | Video or channel url, or a search q
|
The video, the channel or the search results | 7.2 s | Per row: $0.00125 |
| Google Trends |
keyword, geo, timeframe
|
Interest over time with a rising/falling summary, interest by region and related searches | 10.8 s | Per row: $0.00125 |
| LinkedIn Company Pages | Company url
|
Size, website, description, followers and latest posts | 5.3 s | Per row: $0.00175 |
| Instagram Profile | username |
Bio, follower counts, category and latest posts | 11.0 s | Per row: $0.0025 |
| Facebook Pages | Page url
|
Name, categories, contact details, followers and latest post | 5.4 s | Per row: $0.006 |
| Google Search Results |
q, country, language
|
Google's first page of ranked results for the query | 48.9 s | Per row: $0.0015 per results page |
Subreddit or thread url, limit
|
Posts from a subreddit, or a public thread with its comments | 21.8 s | Per row: $0.001 per post, $0.0005 per comment | |
| Mercado Libre Inmuebles |
country, operation, province, limit, or a search url
|
A shortlist of property listings in Argentina, Mexico or Chile | 6.2 s | Per row: $0.0004 per listing |
Per request means the server runs on our Apify account and charges you one event per successful request. Apify notes that these calls don't use your account's memory or concurrent-run limits. Failed requests are not charged. Every per-request price except Zillow's is a new event added on November 18 at the Free-plan price. Paid Apify plans pay 10–30% less.
Per row means the server starts in your own Apify account and bills each result row at the Actor's normal price, just as a regular run does. The Actor page lists the full price, including any start charge. These Actors also save the rows to the run's default dataset.
Copy and paste: three requests
Set your token once. You can find it in Apify Console under Settings → API & Integrations.
export APIFY_TOKEN=your_apify_token
One LinkedIn job by id, billed per request:
curl --get 'https://automatic--linkedin-jobs-scraper.apify.actor/' \
--header "Authorization: Bearer $APIFY_TOKEN" \
--data-urlencode jobId=4431992044
One Zillow home by its zpid, billed per request:
curl --get 'https://automatic--zillow-scraper.apify.actor/' \
--header "Authorization: Bearer $APIFY_TOKEN" \
--data-urlencode zpid=30793376
The latest US Treasury curve, billed per row in your account:
curl --get 'https://automatic--treasury-yield-curve.apify.actor/' \
--header "Authorization: Bearer $APIFY_TOKEN" \
--data-urlencode latest=true
To send the same fields in a JSON body instead, use POST /. That's often easier in n8n or Make. Every response puts the result in data and adds a billing object that says how the request was charged.
Use it from an AI agent, n8n, Make or Zapier
-
AI agents: Give the agent one HTTP tool per Actor. Point it at the Actor's URL, put the Actor's query fields in the tool's parameters, and add the
Authorization: Bearerheader in your code so the model never sees the token. Load/openapi.jsonif your framework builds tools from OpenAPI. Apify also names its MCP server,mcp.apify.com, as a way to reach Server Actors. -
n8n: Add an HTTP Request node with method GET and the Actor URL. Set Authentication to a header credential with
Authorization: Bearer <token>, and add the query fields as query parameters. The node's output is the JSON, so the next node can readdatadirectly. - Make: Use the HTTP → Make a request module with the same URL, header and query string, and turn on Parse response.
-
Zapier: Use a Webhooks by Zapier GET step, or a Code by Zapier step that calls
fetch, with the same header.
When to start a normal run instead
Use the classic Actor run when the job is a list, not a question. Examples are thousands of job ids, every listing in a city, a nightly schedule or a monitor that remembers what it saw last time. That run reads a long input, saves everything to a dataset you can export as CSV or Excel, and sends a webhook when it finishes. If you would otherwise loop over a long list of requests, one run is simpler to manage.
Limits
- Each request returns one object or a short list. Use
limitwhere an Actor offers it, and use a normal run for bulk work. - A server stops after 60 seconds without requests (120 seconds for per-row Actors). The next request starts it again, which adds start-up time to that first answer.
- Every request needs an Apify API token. Wrong or missing fields return HTTP 400 with an
errorobject. Errors are not charged. - Results are the public data the source site shows at request time. If the source site fails or blocks a request, you get an error response.
- Prices are Apify list prices read on November 18, 2026 UTC. Google Search Results moves to $0.002 per results page on November 23, 2026. Mercado Libre Inmuebles changes to tiered listing prices on December 3, 2026. The Actor's Store page always has the current price.
Apify's own announcement is Introducing Server Actors, and the full list from all developers is in the Server Actors collection.
Affiliate disclosure: The Apify Store links carry our referral code. If you sign up for a paid plan through one, ssh.autos may earn a commission at no cost to you.
Originally published at ssh.autos.
Originally published by Dev.to AI. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.