metered-usage-best-practices（计量使用最佳实践）

Metered Usage

Gate access and track consumption with check() and report().

check()

Verify whether a customer can use a feature.

const result = await paykit.check({
  customerId: "user_123",
  featureId: "messages",
})

if (!result.allowed) {
  throw new Error("Usage limit reached")
}

Parameters

Return Value

interface CheckResult {
  allowed: boolean
  balance: {
    limit: number
    remaining: number
    resetAt: Date | null
    unlimited: boolean
  } | null
}

For boolean features: allowed is true/false, balance is null.

For metered features: allowed is true if remaining > 0 (or remaining >= required), balance contains usage details.

Pre-checking Availability

Use required to check if enough units remain before a batch operation:

const { allowed } = await paykit.check({
  customerId: "user_123",
  featureId: "api_calls",
  required: 50,
})

if (!allowed) {
  throw new Error("Not enough API calls remaining")
}

report()

Decrement usage after consumption.

const result = await paykit.report({
  customerId: "user_123",
  featureId: "messages",
  amount: 1,
})

if (!result.success) {
  // Usage limit exceeded
}

Parameters

Return Value

interface ReportResult {
  success: boolean
  balance: {
    limit: number
    remaining: number
    resetAt: Date | null
    unlimited: boolean
  } | null
}

success is false if the customer doesn't have enough remaining balance.

Usage Patterns

Gate before action (check-then-act)

const { allowed } = await paykit.check({
  customerId: userId,
  featureId: "messages",
})

if (!allowed) {
  return { error: "Message limit reached. Upgrade your plan." }
}

await sendMessage(content)

await paykit.report({
  customerId: userId,
  featureId: "messages",
})

Atomic check-and-decrement

For simpler flows, skip check() and use report() directly. It fails if balance is insufficient:

const { success, balance } = await paykit.report({
  customerId: userId,
  featureId: "api_calls",
})

if (!success) {
  return Response.json(
    { error: "API call limit exceeded", resetAt: balance?.resetAt },
    { status: 429 },
  )
}

// Process the API call

Boolean feature gate

const { allowed } = await paykit.check({
  customerId: userId,
  featureId: "custom_branding",
})

if (!allowed) {
  return { error: "Custom branding requires a Pro plan" }
}

Show usage to the user

const { balance } = await paykit.check({
  customerId: userId,
  featureId: "messages",
})

// balance.remaining  // units left
// balance.limit      // total allowed
// balance.resetAt    // when usage resets
// balance.unlimited  // true if no limit

Entitlement Resets

Metered entitlements reset lazily. The reset doesn't happen on a cron. It triggers on the next check() or report() call after the reset time has passed.

Reading Entitlements Directly

Entitlements are also available on the customer object:

const customer = await paykit.getCustomer({ id: "user_123" })

for (const [featureId, entitlement] of Object.entries(customer.entitlements)) {
  console.log(featureId)            // "messages"
  console.log(entitlement.balance)  // current balance
  console.log(entitlement.limit)    // max allowed
  console.log(entitlement.usage)    // consumed
  console.log(entitlement.unlimited) // boolean
  console.log(entitlement.nextResetAt)
}

Complete Example: AI Chat with Usage Limits

// lib/paykit.ts
const messages = feature({ id: "messages", type: "metered" })
const proModels = feature({ id: "pro_models", type: "boolean" })

const free = plan({
  id: "free",
  group: "base",
  default: true,
  includes: [messages({ limit: 50, reset: "day" })],
})

const pro = plan({
  id: "pro",
  group: "base",
  price: { amount: 20, interval: "month" },
  includes: [messages({ limit: 2_000, reset: "day" }), proModels()],
})

// app/api/chat/route.ts
export async function POST(request: Request) {
  const { userId, model, content } = await request.json()

  // Check message quota
  const { allowed, balance } = await paykit.check({
    customerId: userId,
    featureId: "messages",
  })

  if (!allowed) {
    return Response.json({
      error: "Daily message limit reached",
      resetAt: balance?.resetAt,
    }, { status: 429 })
  }

  // Check model access
  if (model === "gpt-4") {
    const { allowed } = await paykit.check({
      customerId: userId,
      featureId: "pro_models",
    })
    if (!allowed) {
      return Response.json(
        { error: "Pro models require a Pro plan" },
        { status: 403 },
      )
    }
  }

  const response = await generateResponse(model, content)

  // Decrement usage
  await paykit.report({
    customerId: userId,
    featureId: "messages",
  })

  return Response.json({ response })
}