Guide
Apify Pricing in 2026: Plans, Compute Units, and Real Costs
Apify's 2026 plans in dollars, what a compute unit buys, the proxy and storage extras, and how the bill compares with per-page pricing.
Apify bills platform usage in dollars, consumed at a compute-unit rate, plus proxies, storage, and data transfer. Two teams running the same number of pages can pay different amounts because the bill tracks memory and run time, not requests. The plan price sets the balance and the rate; the workload decides what those are worth.
The plan prices, included usage, and compute-unit rates below come from Apify's pricing page, read on 15 November 2026. Figures for concurrency and residential proxy rates were checked on 20 October 2026. Pricing changes; re-check the source before committing a budget.
The plans
| Plan | Monthly | Annual (effective monthly) | Included usage per month | Compute-unit rate | Concurrent runs |
|---|---|---|---|---|---|
| Free | $0 | $0 | $5 | $0.20 per CU | 5 |
| Starter | $19 | $17 | $19 | $0.20 per CU | 32 |
| Scale | $199 | $179 | $199 | $0.16 per CU | 128 |
| Business | $999 | $899 | $999 | $0.13 per CU | 256 |
| Enterprise | Custom | Custom | Custom | Custom | Custom |
Annual billing takes 10% off the paid plans. A Creator plan at $1 a month includes a one-time $500 usage allowance for the first six months; it is for people publishing Actors, and publishing on it is restricted to universal Actors.
The included usage is prepaid credit in dollars. It draws from the same balance as compute, proxies, storage, and transfer.
What a compute unit buys
One compute unit is 1 GB of RAM running for one hour. The rate changes across plans: $0.20 per CU on Free and Starter, $0.16 on Scale, $0.13 on Business.
Because the allowance is published in dollars, it converts to compute units at the list rate:
| Plan | Included usage | Compute units at list rate |
|---|---|---|
| Free | $5 | 25 CU |
| Starter | $19 | 95 CU |
| Scale | $199 | 1,244 CU |
| Business | $999 | 7,685 CU |
Those conversions are arithmetic, not spending guidance. Proxy bandwidth and storage draw on the same balance, so the compute units available are lower. Apify does not publish how many pages a compute unit produces; the answer depends on the Actor, its memory setting, and how long the target takes.
Why cost per page varies
The formula explains why two jobs can bill differently:
cost per page = GB of RAM × seconds per page ÷ 3,600 × compute-unit rate
Three arithmetic examples, with all values assumed rather than measured:
| Actor memory | Time per page | CU per page | Cost per 1,000 pages at $0.20/CU | At $0.13/CU |
|---|---|---|---|---|
| 1 GB | 10 seconds | 0.00278 | $0.56 | $0.36 |
| 2 GB | 15 seconds | 0.00833 | $1.67 | $1.08 |
| 4 GB | 30 seconds | 0.03333 | $6.67 | $4.33 |
A memory-heavy Actor that waits on slow pages can cost an order of magnitude more per page than a lean one. Apify charges for the resources the scrape consumes, so a well-tuned Actor is cheap and a poorly tuned one is not.
Residential proxies are billed separately at $7 to $8 per GB depending on the plan. Storage and data transfer draw on the same balance. Some Actors in the Store carry their own rental or per-result pricing set by their authors, so check the Actor page before a large run.
Is Apify free?
The free plan includes $5 of platform usage a month, 16 GB maximum RAM, five concurrent runs, and five datacenter proxy IPs. When the $5 is exhausted, access is blocked until the next billing cycle, and unused usage does not roll over.
The allowance is denominated in compute, not pages. A single small Actor can use several compute units in an hour, so treat the free plan as a short trial.
How Apify's pricing compares with datamint
datamint bills per page at a published multiplier: 1 credit for a static fetch, 5 for a headless render, 15 for stealth routing, and 2 more for a schema pass. Credits are $1 per 1,000, and a failed request costs nothing.
| datamint request | Credits per page | Builder ($49 / 60,000) per 1,000 pages | Scale ($299 / 450,000) per 1,000 pages |
|---|---|---|---|
| Static fetch | 1 | $0.82 | $0.66 |
| Headless render | 5 | $4.08 | $3.32 |
| Stealth routing | 15 | $12.25 | $9.97 |
| Static fetch plus schema | 3 | $2.45 | $1.99 |
The billing units differ, so the comparison depends on what you run:
- For simple jobs, Apify can cost less per page. A 1 GB Actor at 10 seconds per page costs about $0.56 per 1,000 pages at the $0.20 rate, below datamint's $0.82 static rate. Efficient Actors on cheap plans are hard to beat on raw compute.
- datamint is easier to forecast. The cost of a page is published before the run, at 1, 5, or 15 credits plus 2 for a schema pass. There is no memory setting, run-time estimate, or proxy bandwidth line to model.
- Apify's entry price is lower. Starter is $19 against datamint's $49 Builder plan. datamint's free tier is 1,000 credits a month, while Apify's is $5 of usage; the units differ, so neither allowance is directly larger.
- Apify's catalogue is the main reason to choose it. Thousands of prebuilt Actors can cover a common target without writing a scraper. datamint has no Actor marketplace; you supply a URL and, for typed output, a schema.
If your workload is a fixed API call inside an existing pipeline, per-page pricing removes a modelling step. If you want someone else's scraper for a well-known site, Apify's marketplace is the reason to use Apify. The two can coexist as separate services with separate bills.
For the full feature comparison, see Apify alternatives. To price a workload at datamint's rates, see pricing or start with the free tier. For the other credit models in this stack, see the Firecrawl pricing guide and the Browserbase pricing guide.
Frequently asked questions
Is Apify free?
The free plan includes $5 of platform usage a month, 16 GB maximum RAM, five concurrent runs, and five datacenter proxy IPs. It does not roll over, and access is blocked once the allowance is spent until the next cycle.
What is a compute unit?
One compute unit is 1 GB of RAM running for one hour. Apify's rates are $0.20 per CU on Free and Starter, $0.16 on Scale, and $0.13 on Business.
Why is my Apify bill higher than the plan price?
The plan fee is prepaid platform usage, not a cap. Residential proxies, storage, and data transfer draw on the same balance, and usage beyond the allowance is billed at the plan's rate. Unused usage also expires at the end of the cycle rather than rolling over.
Is Apify cheaper than datamint?
For simple jobs on an efficient Actor, Apify can be cheaper per page; one 1 GB Actor at 10 seconds per page costs about $0.56 per 1,000 pages at the $0.20 rate. For fixed API extraction with published per-page credits, datamint is easier to forecast. The two bill different units, so compare at your workload rather than at the headline plan price.
Can datamint replace Apify?
Only for direct URL extraction. datamint has no Actor marketplace and does not run third-party Actors; it is a single API that returns Markdown or schema-validated JSON. If your workflow depends on a prebuilt Actor, there is no equivalent here.
Plan prices, included usage, and compute-unit rates were checked against Apify's pricing page on 15 November 2026; concurrency limits and residential proxy rates were checked on 20 October 2026. Per-page figures are arithmetic from those published rates under stated assumptions, not measured bills. Vendor pricing changes; re-check the source before committing a budget.
datamint returns schema-validated JSON and LLM-ready Markdown from bot-defended sites through one API call, with routing across transport tiers handled on our side.