Skip to main content
AI agents, MCP, and App Store Connect Level: Beginner 10 min read

FlightDeck Licensing & API Quotas: Zero Per-Call Metering Explained

How FlightDeck's local desktop licensing works, why there are zero per-call fees or AI token quotas, and how costs break down across Apple and your AI provider.

Published: 2026-11-03 • Last Updated: 2026-11-03 • Author: Kip Lawrence
Key MCP Tools:

Overview: The Cost Architecture of AI-Driven App Publishing

When developers connect an AI coding assistant (such as Cursor, Claude Code, Windsurf, or Roo Code) to external developer platforms via the Model Context Protocol (MCP), a common financial concern arises:

"Does FlightDeck charge per API request, meter my tool calls, or bill me for AI tokens when my assistant manages App Store Connect?"

The short answer is no. FlightDeck enforces zero per-call fees, zero request metering, and zero token markups.

FlightDeck is an independently licensed native macOS desktop application. It is not an intermediary cloud SaaS or an AI token broker. Understanding how FlightDeck’s local licensing interacts with Apple’s App Store Connect API and your chosen AI model provider enables teams to forecast release operational costs with complete precision.

TEXT
┌─────────────────────────────────────────────────────────────────────────────┐
│                    FLIGHTDECK COST SEPARATION ARCHITECTURE                  │
├─────────────────────────────────────────────────────────────────────────────┤
│ 1. AI MODEL PROVIDER        2. FLIGHTDECK DESKTOP     3. APPLE INC.         │
│ • Anthropic / OpenAI / Google• Flat Desktop License   • $99 / year Program  │
│ • Billed by input/output token• Zero per-call fees    • $0 per API call     │
│ • You pay your AI vendor    • Zero token markup       • Standard rate limits│
│                             • Local MCP on 127.0.0.1  • Direct TLS to Apple │
└─────────────────────────────────────────────────────────────────────────────┘

Operational Planning & Licensing Prompts

Use these prompts with your AI assistant in Cursor, Claude Code, or Windsurf to plan your release tooling economics:

Architecture Cost Audit Prompt

TEXT
I am evaluating FlightDeck MCP for our iOS release management workflow. Audit our technology stack and explain how operational costs are partitioned:
1. Clarify what is covered by our flat FlightDeck desktop license versus our $99/year Apple Developer Program membership.
2. Confirm whether exploratory MCP read calls or bulk metadata operations incur per-call fees or token markups.
3. Outline the network boundaries ensuring our App Store Connect credentials remain local to our workstation.

Three Independent Cost Centers Explained

When using FlightDeck MCP to manage your iOS releases, your financial commitments are split across three completely independent entities. None of these entities take a commission or markup from the others:

1. FlightDeck Desktop Application License

  • What You Pay For: Access to the native macOS FlightDeck desktop software, including the local MCP server, the Desktop Approval Center, the release readiness evaluator, and visual inspection tools.
  • Billing Model: A transparent, flat desktop software subscription (with an initial trial period).
  • Zero Metering: Whether you stage 5 tool calls a month or 5,000 tool calls a day across 50 iOS apps, FlightDeck never bills you for additional requests, extra seats, or binary bandwidth.

2. Apple Developer Program Membership

  • What You Pay For: Active enrollment in the Apple Developer Program to publish apps and access App Store Connect.
  • Billing Model: Exactly $99 USD per year billed directly by Apple Inc. (or $299/year for the Enterprise Program).
  • API Call Pricing: $0.00. Apple does not charge developers for App Store Connect REST API requests. Whether your AI assistant queries build statuses or uploads metadata, Apple provides the API infrastructure at zero incremental cost.

3. AI Model Provider (BYOK — Bring Your Own Key)

  • What You Pay For: The LLM reasoning tokens consumed by your AI editor (Cursor, Claude Code, Windsurf, etc.) while planning operations, formatting release notes, or reading tool outputs.
  • Billing Model: Billed directly by your AI provider (Anthropic, OpenAI, Google) or included in your editor subscription (e.g. Cursor Pro).
  • FlightDeck Involvement: Zero. FlightDeck never proxies your LLM tokens, never resells model access, and never adds a per-token margin.

Why Zero-Metered Architecture Matters for Teams

Many cloud-hosted developer platforms insert themselves as paid relays between your development environment and third-party APIs. This introduces three major pain points that FlightDeck’s architecture deliberately eliminates:

1. No Usage Anxiety During AI Iterations

AI coding agents frequently execute multiple investigative tool calls during a single task. For example, an agent auditing release readiness might call apps_list, app_store_versions_list, app_screenshots_list, and builds_list in quick succession. Under a per-call metering model, developers hesitate to let agents explore because each exploratory call incurs financial overhead. With FlightDeck, tool exploration is unmetered.

2. No Intermediary Markup on Binary Uploads

Uploading a 200 MB .ipa archive through a cloud relay typically incurs outbound bandwidth and compute charges. FlightDeck streams binary chunks directly from your Mac to Apple’s cloud storage over local loopback without passing through third-party cloud infrastructure.

3. Predictable Accounting

Finance teams can budget software expenses without worrying about accidental bill spikes caused by an AI agent getting caught in an automated retry loop.


Rate Limits vs. Billing Quotas: Understanding the Difference

While FlightDeck does not impose billing quotas, developers must distinguish between monetary billing quotas and Apple’s operational infrastructure rate limits:

Dimension FlightDeck Metering Apple App Store Connect API
Per-Call Billing $0 (Zero fees) $0 (Included with Developer Program)
Token Quotas None (Governed by your AI editor) N/A (Apple does not evaluate LLM tokens)
Hourly Rate Limits None imposed by FlightDeck Enforced (Observed ~3,600 req/hour heuristic)
Burst Throttling Built-in Token-Bucket RateLimiter HTTP 429 Too Many Requests

Crucial Operational Caveats:

  1. Free API Access Does Not Mean Infinite Requests: While Apple does not bill for individual API calls, Apple protects its infrastructure using dynamic rate limiting. Excessive concurrent requests or rapid polling will trigger HTTP 429 Too Many Requests.
  2. AI Reasoning Tokens Incur Model Costs: FlightDeck does not charge token fees, but every prompt and tool response consumes tokens billed by your AI assistant's model provider (e.g. Anthropic, OpenAI) or editor plan (e.g. Cursor Pro).
  3. CI/CD Compute is Independent: Cloud compilation runners (such as GitHub Actions macOS runners or Xcode Cloud) bill separately for build minutes. FlightDeck manages App Store Connect release operations; compilation and test runners remain under your own infrastructure contracts.

Telemetry and Privacy Guarantees

Because FlightDeck does not meter your API calls, it has no business incentive or architectural mechanism to inspect your payloads:

  1. Zero Payload Logging: The details of your app metadata, pricing schedules, version descriptions, and screenshots are never logged to external servers.
  2. Local License Verification: Desktop license validation is conducted on-device using secure, anonymous identifiers via the RevenueCat licensing adapter. It never transmits your App Store Connect API keys, bundle IDs, or tool arguments.
  3. Loopback Binding: The FlightDeck MCP server listens strictly on http://127.0.0.1:51122/mcp, preventing external network access to your release tools.

Frequently Asked Questions

Does FlightDeck charge per app managed?

No. You can manage an unlimited number of iOS, iPadOS, macOS, and visionOS applications within your connected Apple Developer accounts under a single desktop license.

What happens if an AI agent makes 100 tool calls in an hour?

All 100 calls are executed at zero additional cost from FlightDeck. The only cost incurred is the token usage charged by your AI assistant's model provider, while Apple handles the API requests within your standard rate limits.

Can FlightDeck shut down my MCP access if I upload too many builds?

No. FlightDeck imposes no volume caps or bandwidth throttling on build uploads. Upload volume is subject only to Apple's standard ingestion policies.


Official References


Connect Your AI Agent to App Store Connect

FlightDeck runs locally on your Mac, keeping private keys secure in your macOS Keychain and executing mutations only after your explicit approval.

Download FlightDeck on Mac App Store