iTechGuides is reader-supported. When you buy through links on our site, we may earn an affiliate commission. As an Amazon Associate I earn from qualifying purchases. Learn more
Pinning Apify Actors can give an MCP agent access to the specific tools you want, but it may also change which pricing and spending-control tools the agent can see. In a field report dated August 12, 2026, Ai-Q Labs for Apify found that its pinned setup omitted pricing discovery and the call-actor wrapper that accepts a run-level charge cap. That is a measurement of one account and configuration, not a guarantee of current behavior for every client. Check the tools your own MCP client actually receives, and put clear pricing and workload limits where the agent can see and enforce them.
What changed when the Actors were pinned?
Apify’s MCP documentation supports selecting tools through URL parameters, including individual Actors. It also recommends explicitly specifying tools for production instead of relying on defaults. Pinning is therefore a legitimate way to make the intended Actors available—but it is worth checking what else the selected tool set includes.
In its August 12, 2026 field report, Ai-Q Labs for Apify compared the tools/list output from its default and pinned configurations. The default listing contained 11 tools. Pinning two of its Actors yielded six tools; pinning all 23 yielded 27. In that account’s pinned listing, the pricing-discovery and caller-side spending-cap tools available in the default listing were absent. The report does not establish that every MCP client, URL, or later server version behaves the same way.
The Tool Desk
Outbyte Driver Updater FREEScan for outdated or missing drivers - takes under a minuteDriver Scan →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →| Configuration in the field report | Tools observed | What the report found |
|---|---|---|
| Default | 11 | Discovery and call tools were available, but the author’s custom Actors were not in the listing. |
| Two Actors pinned | 6 | The selected Actors were exposed; the report says the default pricing-discovery and caller-side cap tools did not survive pinning. |
| All 23 Actors pinned | 27 | The selected Actors were exposed; the report says the same discovery and cap tools were absent. |
These are counts and observations from the author’s setup on that date, not a universal tool-count promise. Tool availability may depend on the URL, server version, and MCP client. Inspect the actual listing after changing the URL rather than assuming pinning preserves every default tool.
Which cost controls are different?
Cost visibility and cost enforcement are separate concerns. An agent needs enough information to estimate what a call might cost, and the call path needs controls that limit billing or the amount of work performed.
Pricing discovery
Apify’s MCP documentation lists search-actors, fetch-actor-details, and call-actor. It says fetch-actor-details can retrieve pricing. If a pinned tool set does not expose pricing discovery, an agent may not be able to look up the current price before invoking an Actor.
Rank #2
Caller-side charge cap
The documented call-actor wrapper accepts maxTotalChargeUsd in callOptions, a maximum total charge for a run through that tool. The field report says its pinned direct Actor tools did not expose that wrapper-level option. A charge cap is not the same as an Actor input limit: it constrains the charge allowed for a run, while input limits constrain what work the Actor is asked to do.
Crashes, No Sound, or Screen Glitches?
Random freezes, missing sound and display glitches usually trace back to one bad driver. Find and replace yours safely.Free scan · under a minuteWindows Errors? Fix Them Before They Spread
Repair common Windows errors and clear accumulated junk for a smoother, more stable PC - no reinstall needed.Free scan · no reinstallRun controls and workload bounds
Apify’s agent onboarding documentation also lists maxItems for Actors priced per result, plus timeout and memory as run controls. These controls address different dimensions of a run and should not be treated as substitutes for bounds in the Actor’s own input schema.
How should you make pinned calls safer?
- Inspect the tools exposed by your exact configuration. Connect the MCP client to the pinned URL, request its current tool list, and confirm whether it includes pricing lookup and the call wrapper with
maxTotalChargeUsd. Repeat this check after changing the URL, client, or server version. - Put pricing information in each pinned tool’s description. State how that specific Actor charges—for example, whether billing is based on an Actor start, dataset items, or another event—and where the agent should verify current pricing. Do not use an old example rate as though it applies to every Actor or remains current.
- Set actual workload bounds in the Actor input schema. Use meaningful JSON Schema constraints such as
maximumon numeric fields that control result counts or other workload. A description that asks the agent to keep a value small is not an enforceable maximum. - Use run-level controls wherever the call path supports them. When calling through
call-actor, set an appropriatemaxTotalChargeUsdand consider the documentedmaxItems,timeout, andmemorycontrols. Verify which of these are available in the tool actually exposed to your client. - Test a bounded call before broad use. Confirm the selected Actor, input limits, and expected billing basis using a small, controlled workload. Review the run’s result and charges before allowing an agent to make larger or repeated calls.
What did the field report’s price and schema checks show?
For its own 23 metered Actors, the report recorded an Actor-start event price of $0.00005 and dataset-item event prices ranging from $0.002 to $0.01, depending on the Actor. It described one run with 608 dataset items and another with 1,000; at the highest per-item rate it recorded, 1,000 item events would amount to $10. These are the report author’s observed rates and run records, not current or universal Apify prices. Check the current details for the Actor you intend to call.
The author also found that all 23 Actors had input fields intended to bound work, but none declared a JSON Schema maximum for those fields. That finding applies only to the Actors reviewed in the report. The practical lesson is to make limits enforceable in the schema rather than relying on an agent to infer a safe value from a description.
What the report does—and does not—establish
The report is a dated first-person measurement, not an independent replication or a guarantee of how the service behaves today. Its author also reported a median of per-Actor medians of 4.05 seconds, with only five of the 23 Actors having enough history to calculate a median, and a slowest observed run of 58.0 seconds. Those account-specific latency figures are not a general performance benchmark and do not predict the time or cost of another Actor’s run.
Recommended Free Tools
The Apify MCP server repository describes payment options including AGI, which provides a prepaid, spend-capped API token usable for any Actor, and direct x402 and Skyfire payments, which it describes as limited to Pay Per Event Actors. The repository does not say these options restore per-call pricing visibility in pinned tool descriptions; payment choice and agent-visible cost information remain distinct questions.
Quick Recap
Best Value
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
