Your AI agent does the wiring over MCP. You add one call. Here's the whole path.
Sign in and create a project — name it and pick your platform. You'll get a server-side API key. It stays on your backend / in your agent's MCP config; it's never shipped in the app.
Design packages, pricing, and copy in the visual builder with a live phone preview. The config is saved server-side and fetched by the SDK at runtime — change it later without an app release.
Point Claude Code, Codex, or Cursor at the Tierux MCP server with your key. Then describe what you want — the agent maps your store products, wires the SDK, and returns a preview to approve before anything is saved.
// .mcp.json { "mcpServers": { "tierux": { "url": "https://mcp-tierux.web.app", "headers": { "Authorization": "Bearer spe_live_…" } } } }
▸ "Set up a Pro monthly sub with a 7-day trial and wire it into my app." setup_google_play() create_play_product(id: "pro_monthly", trial: "P7D") create_product_mapping(product: "pro_monthly", entitlement: "pro") generate_sdk_snippet(platform: "android")
Then it's one call in your app:
// Android (Kotlin) Tierux.showPaywall("pro") // later, anywhere — gate a feature if (Tierux.isEntitled("pro")) { unlockProFeatures() }
Run a test purchase. showPaywall renders your remote-configured paywall, runs the native store purchase, and Tierux verifies it server-side — normalizing it into queryable entitlement state, with the verification showing live in your dashboard. Raw purchase tokens are never stored.
Verification is server-side. Your app never carries a secret that could be extracted from the binary.
Change packages, pricing copy, and layout without an app release — the SDK fetches the live config.
Same flow on both. Android is the native core; the iOS Swift SDK + App Store verification ship today.