You need listings or prices from a site, and you don’t want to write a scraper.
To use Apify Actors without code: pick a Store Actor, fill the form, run it, export CSV or JSON. Free includes $5 of usage a month.
We checked the steps against Apify’s docs and pricing mid-October. No paid run for this guide.
Skip compute units and that $5 can vanish before lunch.
What is an Apify Actor?
Disclosure: some Apify links are affiliate links. If you upgrade through them, we may earn a commission. The steps and the warnings stay the same either way.
An Apify Actor is a serverless program that takes JSON input, does a job (scrape a site, drive a browser, clean a dataset), and writes results to storage you can download later.
You run it in the Apify Console, through the API or CLI, or on a schedule. Public Actors live in Apify Store. Private ones stay on your account.
Our web scraping tools roundup places Apify next to API scrapers and no-code tools. This post is the hands-on path.
What do you need before your first run?
- An Apify account. Free works. Card optional until you outgrow the $5 monthly usage credit.
- A clear target. One URL pattern or search query beats “scrape the whole internet”.
- A low first limit. Cap pages or results at 10–50 until the columns look right.
- Five minutes to read the Actor README and Pricing tab. That read saves more money than any memory tweak.
No local install. No Python. The Console is enough for this tutorial.
How do you pick the right Actor from the Store?
Search Apify Store for the site or the job. Open a few candidates and score them on four things:
- README clarity. Does it show required inputs and sample output?
- Pricing model. Pay per event (per result, per start) vs pay per usage (compute + proxy + storage).
- Recent runs. A popular Actor with fresh success history beats an abandoned one with a flashy title.
- Memory hint. Browser Actors need at least 1024 MB. Cheerio-style HTML scrapers often run fine at 1024–2048 MB.
For cost maths across plans and Actor types, keep our Apify pricing breakdown open in another tab while you decide.
Run your first Actor in the Console
The loop never changes: pick → input → start → collect. Here’s the Console path.
Run your first Apify Actor
Store pick, input, start, export
-
Step 1 of 5
Create a free Apify account
Sign up on Apify. The Free plan gives you $5 of platform usage a month and needs no card. Stay on it until a real job outgrows it.
-
Step 2 of 5
Open Apify Store and search your site
Search for the site or job (Google Maps, Instagram, e-commerce, sitemap crawl). Open 2–3 Actors and compare the README, recent success rate, and the Pricing tab.
-
Step 3 of 5
Read the Pricing tab before you start
Note whether it is pay per event or pay per usage. A cheap-looking Actor that bills per result can still burn residential proxy gigabytes on top.
-
Step 4 of 5
Fill the input form
Enter start URLs or search terms, set a low max-items or max-pages limit for the first run, and leave memory at the Actor default unless the README says otherwise.
-
Step 5 of 5
Start the run and open the Dataset
Click Start. When the run finishes (or while it runs), open the Dataset tab. Export as CSV, JSON or Excel, or copy the API endpoint for later.
That form is JSON under the hood, the same object the API and schedule overrides use. You don’t need JSON on day one.
Input options that actually matter
- Build. Leave on
latestunless the README pins a version. - Timeout. Seconds until Apify kills the run.
0means no timeout. Set a real timeout on schedules so a stuck job can’t run all night. - Memory. Power of 2 only: 128, 256, 512, 1024, 2048, 4096 MB and up. Higher memory buys more CPU (about one core per 4096 MB).
Our take: first run at the Actor’s default memory with a tiny result limit. Raise memory only if the run is slow and the README says concurrency will scale with RAM.
How do you export the results?
Structured rows land in the run’s Dataset. Files and screenshots go to the Key-value store. Always check the README for which one the Actor fills.
From the run page:
- Open Dataset → Export → CSV, JSON or Excel.
- Or copy the Dataset API URL if another tool will pull the rows.
- Open Storage when you need every related store (default dataset, key-value store, request queue).
Dataset rows stay after the run ends, so you can download again later. Storage reads still count toward usage, so don’t poll a huge dataset every minute.
Save a Task, then schedule it
Once a manual run looks right, stop retyping the form. Save a Task, then put that Task on a schedule.
Turn a good run into a schedule
Task first, then cron
-
Step 1 of 4
Save the input as a Task
On the Actor page, save your working input as a named Task. Next time you open that Task, every field is already filled.
-
Step 2 of 4
Confirm you have run it at least once
Apify only lets you schedule an Actor after one successful manual run. Use that run to prove the input and the export look right.
-
Step 3 of 4
Create a Schedule and set timezone
In Console → Schedules → Create new. Name it, pick your timezone (DST is supported), and set the cron. Example:
0 8 * * *runs daily at 08:00. -
Step 4 of 4
Attach the Task and enable
Add your Task (or Actor). New schedules start disabled. Enable only after you check the next fire time. One schedule can hold up to 10 Actors and 10 Tasks.
Cron cheatsheet for Schedules (minute hour day month weekday):
# minute hour day-of-month month day-of-week
0 8 * * * # every day at 08:00
0 */6 * * * # every 6 hours
0 9 * * 1-5 # weekdays at 09:00
30 2 1 * * # 02:30 on the 1st of each monthOne schedule can fire several Actors. Keep notifications on until you trust the setup; Apify emails you if a Task fails to start.
What is a compute unit, in plain English?
A compute unit (CU) is Apify’s bill for CPU and RAM over time.
1 CU = 1 GB of memory running for 1 hour.
Worked examples you can reuse:
| Memory | Duration | CU maths | CU total |
|---|---|---|---|
| 1024 MB (1 GB) | 1 hour | 1 × 1 | 1 CU |
| 4096 MB (4 GB) | 15 minutes | 4 × 0.25 | 1 CU |
| 2048 MB (2 GB) | 30 minutes | 2 × 0.5 | 1 CU |
| 1024 MB | 6 minutes | 1 × 0.1 | 0.1 CU |
On Free and Starter, platform compute is billed at $0.20 per CU. Scale drops that to $0.16, Business to $0.13 (pricing checked mid-October). Your plan also includes a prepaid usage pot ($5 Free, $19 Starter, $199 Scale) that pays for CUs, proxies, storage ops and data transfer until it runs out.
CU is only one line on the bill. Proxy traffic and pay-per-event fees often dwarf compute on hard sites. See our Apify pricing guide for the full tables.
Memory mistakes that waste CU
- Browser Actors under 1024 MB. Puppeteer and Playwright need at least 1 GB. Starving them makes runs fail or crawl.
- Cheerio Actors at 16 GB. Plain HTML scrapers rarely use more than one CPU core. Past about 4096 MB you often pay for RAM you never use.
- Dozens of tiny runs. Each start has overhead. One longer run with a sane limit usually costs less CU than fifty five-second runs.
- Ignoring autoscaling. On multi-URL Crawlee Actors, doubling memory can cut duration roughly in half, so CU stays similar while you finish sooner.
Pay per event vs pay per usage
Every Store Actor states its model on the Pricing tab.
- Pay per event. You pay fixed prices for events the developer defines (Actor start, each result, each enriched row). Many PPE Actors fold platform usage into the event price. Some still add CU or proxy on top. Read the fine print.
- Pay per usage. The developer adds no markup. You only pay Apify platform usage: compute units, transfer, storage, proxies.
Test with a tiny limit either way. A 20-result run beats any marketing table.
Five mistakes that empty your usage credit
Skipping the Pricing tab. A “popular” Actor can charge per result plus residential proxy. Your $5 Free pot disappears on one wide Maps scrape.
Leaving max items uncapped. Defaults can mean thousands of pages. Set a hard cap on the first three runs.
Scheduling before the export looks right. A daily cron that writes empty rows is an expensive alarm clock.
Timeout = 0 on a flaky site. No timeout means a hung browser can burn CU until you notice.
Treating public data as a free-for-all. Public pages still have terms, and personal data has extra rules. Scrape only what you’re allowed to keep.
When should you use Apify instead of Firecrawl?
Pick Apify Actors when you want a ready-made scraper for a named platform, a Console workflow with Tasks and Schedules, or storage that keeps datasets between runs.
Pick Firecrawl when the job is “turn these URLs into clean Markdown or JSON for an AI app” and you’re happy calling an API from code or MCP.
We compared cost and fit in Firecrawl vs Apify, and added a third option in Firecrawl vs Cloro vs Apify. Credit maths for Firecrawl alone live on our Firecrawl pricing page.
If the hard part is “browser or HTTP?”, our browser automation vs web scraping note explains why a Playwright Actor costs more CU than a Cheerio one for the same URL list.
Bottom line
To use Apify Actors without writing code: prove one tiny Store run, export the Dataset, save a Task, then schedule it.
Watch the Pricing tab and the compute unit maths before you scale. Free’s $5 is enough to learn the loop. Paid plans only make sense once a Task earns its keep.
Frequently asked questions
Do I need to know how to code to use Apify Actors?
No. Apify Store Actors run from the Console with a form for inputs. You can export CSV or Excel when the run finishes. Code and the API help later for automation, but they are optional for the first successful scrape.
What is the difference between an Actor and a Task?
An Actor is the tool. A Task is that tool with your saved input and run options. Save a Task once the manual run looks right, then re-run or schedule the Task without filling the form again.
How do compute units relate to my Free plan?
Free includes $5 of platform usage per month. Compute is one part of that pot, billed around $0.20 per CU on Free, alongside proxies, storage and transfer. When the $5 is gone, Free access pauses until the next billing cycle.
Can I schedule an Actor I have never run?
No. Apify expects at least one manual run before you attach an Actor to a Schedule. That first run proves the input works and gives you a Dataset to check.
Should I always raise memory to finish faster?
Not always. Browser Actors need at least 1024 MB. Autoscaling crawlers can finish sooner at higher RAM with similar CU. Single-page or Cheerio jobs often waste money above 4096 MB because Node will not use the extra cores.
Where do screenshots and files go if the Dataset looks empty?
Check the Key-value store on the run Storage tab. Many Actors put tabular rows in the Dataset and put files, HTML snapshots or screenshots in the key-value store instead. The README usually says which store to open first.
Is pay per event cheaper than pay per usage?
It depends on the Actor and the site. Pay per event is easier to forecast per result. Pay per usage can be cheaper on light HTML jobs and painful on proxy-heavy browser crawls. Run a 20-result test and read the Usage breakdown before you scale.
Can I send Actor results into n8n or Make?
Yes. Apify supports webhooks and integrations so a finished run can notify Slack or push data into automation tools. Get a clean Dataset export first, then add the webhook so you are not debugging two systems at once.
