Browse documentation
Last reviewed

FAQ

AdPilot MCP connects authorized AI clients to mapped Apple Ads and optional RevenueCat data. The answers below cover the setup, measurement, security, limits, and operational behaviors that most teams need first.

Is AdPilot MCP another ads dashboard?

No. The SaaS control panel is for configuration, limits, skills, and detailed audit history. Analysis and conversation happen in your AI client over MCP.

Which AI clients work?

Any client that supports remote MCP with OAuth—commonly ChatGPT, Claude, Cursor, and Codex. See Connect your AI client.

Does the assistant change my campaigns automatically?

Mutation tools execute immediately when the subscription permits writes and workspace limits and correctness rules pass. Every operation returns an action ID and is visible in Changes.

Scheduled email reports are different: they are always read-only. They email recommendations to the workspace owner and never change Apple Ads. Use a separate direct MCP request when you want to make a change.

How much does AdPilot cost?

Free supports the owner only, one app, one MCP client, 30 days of history, and read-only analysis. Solo includes the owner plus one teammate for $29 per month or $290 per year. Studio includes the owner plus up to four teammates for $99 per month or $990 per year. Workspace owners can upgrade through Lemon Squeezy checkout and manage an existing subscription from Billing settings. See Plans, pricing, and subscription limits.

Can I invite teammates to my workspace?

Yes, when the plan includes an available member seat. Free is owner-only, Solo lets the owner invite one teammate, and Studio supports up to four teammates. Owners and Admins send invitations from Settings → Members. Active pending invitations reserve seats and expire after seven days. See Invite and manage workspace members.

Which workspace role should I assign?

Use Viewer for read-only access, Operator for strategy and permitted Apple Ads operations, and Admin for trusted teammates who must manage workspace configuration and members. Only the owner controls billing, deletion, and ownership transfer, and only the owner can grant Admin access.

Does the AdPilot price include Apple Ads spend or my AI subscription?

No. Apple bills campaign spend directly through the advertiser’s Apple Ads account. Your connected ChatGPT, Claude, Cursor, Codex, or other MCP client may also require its own plan or usage payment. AdPilot’s paid prices are flat workspace subscriptions and do not take a percentage of Apple Ads spend.

Why was my scheduled report skipped?

Each workspace has a monthly report budget. If the next occurrence cannot reserve budget, AdPilot records a skipped run without making an AI request or sending email. The schedule stays enabled and tries again at its next normal weekly or monthly occurrence after the monthly reset.

Why do I see “complete days”?

Metrics tools exclude the incomplete current day so comparisons stay honest across timezones and sync lag.

Do I need RevenueCat?

Apple Ads-only analysis works without it. Profitability and proceeds questions need RevenueCat connected and mapped. AdPilot reports blended app-level proceeds and does not claim deterministic campaign-level RevenueCat attribution. See RevenueCat and profitability.

Do I need a separate Apple Account for the API?

A dedicated, customer-owned Apple Account is the recommended production API identity. First set up Apple Ads under the correct individual or legal entity and retain an Account Admin. The admin then invites the API identity, which accepts with the same email address. Create the API client later during the AdPilot connection flow. AdPilot never needs the Apple password or two-factor authentication code. See Set up an Apple Ads account and API user.

Which Apple Ads API role should I choose?

Use API Account Read Only for analysis-only access and API Account Manager when supported direct operations should work across the account. Apple also offers API Read Only and API Read & Write roles limited to selected campaign groups. Use the least privilege that matches your intended workflow.

Does AdPilot support Apple's service-provider Sign in with Apple flow?

No. AdPilot supports customer-managed API credentials. You or your organization creates and controls the Apple Ads API user and client, registers the AdPilot-generated public key, and enters the Client, Team, and Key IDs in AdPilot. AdPilot does not currently offer Apple's service-provider Sign in with Apple authorization flow. See Connect Apple Ads to AdPilot MCP and Apple's Campaign Management API guidance.

What is the difference between CPI and ROAS?

CPI divides spend by attributed installs; it measures acquisition cost. ROAS divides observed or attributed value by spend; it measures value recovery. A low CPI can still produce weak ROAS if users do not pay or retain. Read Apple Ads metrics and formulas for formulas and zero-denominator rules.

Where do private keys live?

Apple Ads private keys are generated and encrypted in AdPilot. They are not shown in the UI after creation and are never returned over MCP. You upload the public key to Apple.

What if an operation is blocked?

A failed operation retains its action ID and structured reason—for example a protected term, bid over the cap, spend projection over the ceiling, or provider rejection. Fix the underlying limit or data issue, then retry.

Why is month-to-date $0 but the projected month is much higher?

Month-to-date uses synchronized actual spend from complete days. On the first day of a month, that can correctly be zero. The projection also includes the maximum daily-budget exposure for enabled campaigns through month-end. See Spend limits and projections.

Why did changing app monetization not change the projection?

Monetization model, trial duration, notes, and RevenueCat data guide profitability analysis but do not change Apple campaign exposure. Campaign statuses and budgets change the projection. A confirmed app monthly ad-spend cap changes what MCP may authorize, but it does not change the projected amount itself.

Does Save draft activate my app monthly cap?

No. Save draft stores unfinished strategy edits. Use Confirm strategy to activate the app cap for subsequent MCP operations.

Can my ceiling be below the current projection?

Yes. AdPilot does not automatically pause campaigns when you lower a ceiling. It blocks operations that would increase exposure while continuing to allow pauses, unchanged budget values, and net budget reductions.

Should I raise the ceiling when an operation is blocked?

Only if the higher maximum exposure is intentional. Raising the ceiling expands what MCP may authorize. If the current projection is already too high, reduce enabled campaign budgets or pause campaigns instead.

Can I undo a change?

Reversible applied actions offer Rollback from Changes. Not every Apple mutation is reversible; the UI indicates when rollback is available.

How do skills differ from tools?

Tools are individual MCP operations (list campaigns, prepare a negative, get ROAS). Skills are ready-to-use playbooks that tell the assistant which tools to use and in what order. Explore them on the Skills page.

Who can read these docs?

Documentation is public so you can share setup guides before or after signup. Workspace data, keys, and MCP access always require authentication.

How do I contact support?

Email support@kovacssoftware.com. For account, connection, report, or campaign-operation issues, include the workspace name and any relevant action or run ID. Do not send Apple Ads private keys, passwords, or two-factor authentication codes.

Where can I learn more about limits and action history?

Read Direct operations, spend limits, and audit for formulas, examples, error codes, rollback, and troubleshooting.

Quick troubleshooting checklist

  1. Confirm the Apple connection and metrics sync are current.
  2. Confirm the campaign is mapped to the intended app.
  3. Use complete days in the workspace timezone.
  4. Check RevenueCat mapping, currency, and cohort maturity before interpreting ROAS.
  5. Copy the action ID and structured error before retrying a blocked change.

How AdPilot uses these rules

AdPilot applies these behaviors in its server-side data, authorization, limit, and audit services. The AI client can explain a result or request an operation, but it cannot override missing data, workspace permissions, protected terms, or configured spend controls.

Next steps

Official sources

AdPilot is an independent service and is not affiliated with, endorsed by, sponsored by, or operated by Apple Inc. Apple and Apple Ads are trademarks of Apple Inc.