$0 Wins: Apify's Free Tier Beats Paid 1.4× Per Actor
FREE-pricing Apify actors (6.1% of tracked catalog) capture 11.1% of demand: 233 users/actor vs 168 for PPE and 9.5 for Rental. The $0 tier out-densifies paid by 1.4×. Half are OPEN_SOURCE; the other half are publisher monetization plays that abandoned paywall friction.
Of 3,155 Apify actors with continuous data history, the 6.1% on FREE pricing capture 11.1% of measured demand — 44,982 of 405,438 monthly active users. The per-actor numbers are sharper: FREE actors average 233 users/month, against 168 for pay-per-event and 9.5 for the dying Rental tier. The $0 tier punches 1.4× its weight against paid models on a per-actor basis.
Publishers who treat FREE as a signal of weakness have the math backwards. On the Apify Store, removing the price tag is correlated with winning more demand than the median paid actor — not less.
The $0 tier outperforms
The pricing-model split across the tracked subset:
| Model | Actors | Catalog share | Users/30d | Demand share | Users/actor |
|---|---|---|---|---|---|
| PAY_PER_EVENT | 2,093 | 66.3% | 352,202 | 86.9% | 168 |
| FREE | 193 | 6.1% | 44,982 | 11.1% | 233 |
| FLAT_PRICE_PER_MONTH (Rental) | 869 | 27.5% | 8,254 | 2.0% | 9.5 |
The Rental tier — Apify’s legacy subscription pricing — is retiring on October 1, 2026. The 2.0% demand share on 27.5% of the tracked catalog tells the story. The decision to deprecate did not get made in product strategy meetings. It got made on the demand chart.
What is harder to read at first pass is the FREE-vs-PPE comparison. PPE is the platform default, the recommended migration target for retiring Rental publishers, and the model that captures 86.9% of demand. But on a per-actor basis, FREE actors are out-densifying PPE by 1.4×. The same demand pool, divided across one-tenth as many actors, produces a per-actor reward that the modal PPE publisher does not see.
Where the FREE actors come from
Two cohorts produce the 193 FREE actors, and they do not overlap as much as the language might suggest.
The OPEN_SOURCE intersection. Of 47 OPEN_SOURCE-tagged actors on the tracked subset, 24 (51.1%) are priced at FREE. The other half are commercialized open-source — the upstream project is OSS, but the actor wrapper around it carries a price. The 24 FREE-and-OPEN_SOURCE actors are the canonical “publishers giving away their work” cohort.
The monetization-abandoner cohort. The remaining 169 FREE actors (87.6% of the FREE pool) are not OPEN_SOURCE-tagged. These are publishers who chose to remove the paywall as a positioning move — typically to maximize adoption, to feed a downstream paid product, or because the actor is a loss-leader for the creator’s portfolio. This cohort outnumbers the OSS-and-FREE actors by 7:1.
The aggregate density advantage holds across both. FREE-and-OPEN_SOURCE has even higher per-actor demand than FREE-not-OS, but both bands beat the PPE per-actor median.
What Rental’s collapse confirms
The 9.5 users-per-actor density on FLAT_PRICE_PER_MONTH is not a measurement artifact. The Rental tier has 869 actors and 8,254 monthly active users. Most Rental actors are sub-10 MAU. The cohort that has not yet migrated to PPE is concentrated in the long tail of the catalog, where the subscription price-point ($30-100/month for typical Rental actors) is mismatched with the volume the actor can serve.
Apify’s retirement of Rental on October 1, 2026 forces this cohort into one of two moves: migrate to PPE (where the per-actor demand floor is higher) or migrate to FREE (where the density is highest of all). Publishers who do nothing become unfindable on October 2. The platform-design pressure points the long tail at PPE; the demand-density data points it at FREE.
The non-trivial question for the 869 affected publishers: if their actor cannot sustain a $30/month subscription audience, can it sustain a per-event price floor? For most, the honest answer is no. The migration path that actually preserves demand is to FREE — accept the zero unit-economics and reposition for adoption.
What free pricing signals to the MCP-era buyer
The implication for the MCP-era buyer — an LLM agent picking an actor from a typed tool list — sharpens the FREE-vs-paid economics further.
An LLM agent reads the tool’s input schema, output schema, and price tag at runtime. A FREE tool with comparable success rate to a paid alternative is the agent’s correct selection nine times out of ten. There is no brand loyalty in the tool-selection loop. There is only price, latency, and observed success.
That dynamic explains part of what the demand-density data is already showing. The 233 users-per-FREE-actor figure is partly humans choosing free over paid for trivial reasons (no signup friction), and partly agents and orchestration tools choosing free over paid for cost-arbitrage reasons. The agent-driven share of that 233 is small now and growing.
For Apify Store publishers in the long tail, the FREE tier looks like a defensive move (give up the paywall to stay alive). For publishers in the high-density categories — OPEN_SOURCE, TRAVEL, SOCIAL_MEDIA — FREE looks like an offensive move (capture the cost-arbitrage flow that paid competitors cede). The pricing decision is the same; the strategic context is opposite.
By Q4 2026, when Rental has fully retired and the FREE cohort has absorbed some of the displaced publishers, the per-actor density gap between FREE and PPE will be the most consequential pricing question on the Apify Store. The number to watch is not the catalog share. It is the demand-per-actor on each tier.
Sources
- Apify pricing models documentation
- Apify Store — pricing model selector
- Signal Census pulse data — 3,155 actors with continuous data history as of 2026-05-16
- Signal Census: Apify Store May 2026 — full headline pricing-mix breakdown
- Signal Census: 3-Tag Default Backfires — companion analysis on tag discipline