Apify Review: Is the Actor Platform Worth It?

2. What you are actually buying

Apify sells a platform for running Actors — store scrapers and your own jobs — not a single “HTML in, Markdown out” credit pack. Three meters show up on almost every invoice, and a fourth shows up on most store actors:

3. Official plans and side meters

Annual billing is 10% off. The prepaid amount equals the listed monthly price on each paid tier. Residential proxy list prices (separate from actor events): $8/GB on Free and Starter, $7.50 on Scale, $7 on Business. Datacenter IPs: 5 included on Free; 30 included then $1/IP on Starter. A paid add-on is listed at $5 per extra concurrent run and $1 per extra GB of actor RAM.

CU capacity if a job were compute-only, using the published rates: Free’s $5 buys 25 CU (25 GB-hours) at $0.20. Starter’s $19 prepaid buys 95 CU; with the public “at most $100” overage description the ceiling is $19 + $100 = $119, or 595 CU at $0.20. Scale’s $199 prepaid buys 1,243.75 CU at $0.16. Business’s $999 prepaid buys about 7,684 CU at $0.13. That Starter ceiling matters: a 1,000-CU compute month ($200 at $0.20) cannot fit the public overage cap, so the next documented home is Scale, not “Starter plus a large overage.”

4. Real cost by scenario

Scenario A — The official Instagram example

The vendor’s published rate is $2.30 per 1,000 results.

• 1,000 results = $2.30, which fits Free’s $5 (Free holds about 2,170 results at that rate if nothing else bills).
• 10,000 results = $23. That overflows Starter’s $19 prepaid slice by $4, billed as overage if you have set a usage limit that allows it.
• 50,000 results = $115. Overage against Starter would be $96, under the public $100 cap, for a total of $115.
• 60,000 results = $138. Overage would be $119, which is above the public “at most $100” Starter description — this volume does not fit Starter as publicly described. Scale’s $199 prepaid holds it, and at $2.30 / 1,000 results Scale’s wallet covers about 86,500 results.

Those numbers assume the actor bills only the published event price. If the same run also pulls residential GB or compute, the invoice rises on those meters.

Scenario B — A 10-hour job at 2 GB of RAM

Official CU definition: 10 hours × 2 GB = 20 CU.

• Free or Starter: 20 × $0.20 = $4.
• Scale: 20 × $0.16 = $3.20.
• Business: 20 × $0.13 = $2.60.

The $0.80 you save by running that one job on Scale instead of Starter does not justify Scale’s $180 sticker gap. Scale starts to make compute-only sense when Starter cannot hold the month. A daily two-hour job at 2 GB is 4 CU per day, or 120 CU in a 30-day month: $24 at Starter’s $0.20 (so $19 prepaid + $5 overage) versus $19.20 of Scale’s $199 wallet. Stay on Starter. A 1,000-CU compute month is $200 at $0.20 and does not fit Starter’s public overage cap; it fits Scale at $160 of the $199 included.

Scenario C — 2 GB of residential proxy on Starter

Residential list price on Free and Starter is $8/GB, so 2 GB = $16 — almost the entire $19 prepaid slice — before any actor event or CU line. Five GB is $40, which is already into Starter overage ($19 + $21). The same 5 GB on Scale is 5 × $7.50 = $37.50 of a $199 wallet. If the reason you opened Apify was “I need residential GB,” you are shopping the wrong aisle; compare Bright Data or the Bright Data vs Oxylabs pairwise rather than treating Apify’s proxy table as the product.

Scenario D — Concurrency and extra RAM

Free allows 5 concurrent runs; Starter 32; Scale 128; Business 256. An extra concurrent run is listed at $5; an extra GB of actor RAM at $1. A morning fan-out that already queues at 5 workers is a Starter reason (32), not evidence the platform failed. Paying $5 × 10 extra runs as a habit, on top of Starter, is a signal to read Scale’s 128-run cap — but only after you price the usage those runs will burn.

Scenario E — Creator versus Starter

Creator is $1/month with a one-time $500 of platform usage for the first six months, and the Store is limited to universal Actors. That is a builder SKU. It is not Starter at a discount, it is not 26 months of Starter’s $19 prepaid, and it does not replace Starter if you need the full store or production concurrency. When the six-month $500 is gone, you are not looking at a hidden production plan — you are looking at the public ladder again.

Scenario F — Prototype week on Free

Free’s $5 is a sandbox: about 2,170 Instagram-example results, or 25 CU, and then the plan blocks until the next cycle. It is large enough to read an actor page and run a named job once. It is not a production budget, and emptying it is not evidence you should leave the category — it is evidence you should read that actor’s price and, if actors still match the job, move to Starter.

5. How to estimate a bill

1. Pick or build the actor first, then read its pricing model on the actor page (event, usage, or both). The $19 Starter fee is not the monthly invoice until you know those lines. We will not invent CU burn for a crawler we did not run.
2. Convert the published event price at your volume, the way the Instagram example converts: results ÷ 1,000 × $2.30, or whatever that actor page prints.
3. Add compute if the actor is usage-priced or mixed: GB of RAM × hours = CU, then multiply by $0.20 / $0.16 / $0.13 depending on the plan.
4. Add proxy if the run will use it. Residential GB at $8 on Starter can outrun the prepaid slice on a job that looked “simple” on the event line.
5. Set a usage limit (Starter overage publicly at most $100) and compare the total to Free $5, Starter $19 plus that cap, and Scale $199.
6. Check concurrency (5 / 32 / 128 / 256) separately from dollars. Queueing at 5 workers is a plan step, not a category change.
7. Only then look at Creator, and only if you are building and the universal-Actor Store limit is acceptable.

6. Pitfalls

• Treating Starter’s $19 as a scrape-API credit pack or as Firecrawl Hobby’s 5,000 credits.
• Skipping the actor page and averaging the store. Event prices are per actor; the Instagram $2.30 figure is an example, not a platform average.
• Forgetting that unused prepaid usage expires at cycle end — a light month still costs $19 on Starter, and leftover dollars do not roll.
• Assuming Starter can absorb any overage. The public description caps the overage you set at $100, which at $0.20/CU is 500 CU beyond the prepaid 95.
• Buying Apify to “do RAG cheaper” without counting event prices plus residential proxy. If the output you wanted was Markdown, you started in the wrong category.
• Using the Creator plan as if it were Starter.
• Ignoring concurrency (5 / 32 / 128 / 256) when a morning fan-out already queues.
• Jumping to Scale to save $0.04 per CU on a 20-CU job ($0.80) while paying a $180 sticker gap.

7. Decision

Apify is worth it for teams whose job is running actors, not fetching pages through a single credit weight. We have no affiliate link to offer today; a hidden placeholder sits in the HTML for later.

8. FAQ

Is Apify worth it?

Yes, if you need a store or custom actors and you will read each actor’s price. No, if you wanted a hosted Markdown API or a thin scrape API. Official paid entry is Starter at $19/month ($17 annual display) with $19 of included usage, checked November 6, 2026.

What is an Apify compute unit?

1 CU = 1 GB of RAM running for one hour. Official prices: $0.20 on Free and Starter, $0.16 on Scale, $0.13 on Business. A 10-hour job at 2 GB is 20 CU, or $4 on Starter.

Does unused usage roll over?

No. Official materials: unused prepaid usage expires at the end of the billing cycle.

How far will Free’s $5 go?

About 2,170 results at the official Instagram example of $2.30 per 1,000, or 25 CU at $0.20. Then the free plan blocks until the next cycle.

Is the Creator plan a cheaper Starter?

No. It is $1/month with a one-time $500 of platform usage for the first six months, and the Store is limited to universal Actors. Treat it as a builder SKU.

Do you earn a commission on Apify?

Not at the time this page was written. A hidden placeholder sits in the HTML for later; it is not visible and does not point at a live vendor URL.

9. Sources

• Apify official pricing page, including plan table, CU prices, proxy table, actor-pricing FAQ, and worked examples (checked November 6, 2026)
• Apify official creator-plan page as cited from the pricing page (checked November 6, 2026)