Can agents use Stripe?

What the agent told the user

**Yes, Stripe directly supports usage-based billing for AI products charging per token.** According to Stripe’s documentation, there are **two ways** to implement per-token usage-based billing on Stripe: 1. **Metronome (Recommended for new integrations):** Stripe’s primary usage-based billing platform, which now includes a dedicated **Bill for LLM tokens** feature (in public preview). 2. **Basic Usage-Based Billing (Stripe Billing Meters API):** Stripe’s lower-level native metering primitive, which remains fully supported (and offers direct compatibility with Stripe Connect, Checkout, Adaptive Pricing, and Workflows). Here is how both approaches work according to the Stripe documentation: --- ### Option 1: Metronome & "Bill for LLM Tokens" (Recommended for New Integrations) Metronome (now part of Stripe) is Stripe’s primary platform for real-time metering, pricing, and billing, and is recommended when you need high-volume event ingestion, real-time usage visibility, prepaid credit burndown, or dimensional pricing (e.g., pricing by model and token type). #### How It Works 1. **Choose Your Token Billing Approach:** * **Charge your app’s users:** If you build an AI app and charge end users for token consumption, API calls, or compute time. * **Resell LLM tokens:** If you run an AI gateway or platform providing access to third-party LLM providers and want to set a markup/margin on token costs. Stripe automatically syncs model prices across **OpenAI, Anthropic, and Google** models so your pricing updates automatically when providers change rates or release new models. 2. **Define Your Pricing Model in Metronome:** * **Pure pay-as-you-go:** Bill strictly per token consumed with no fixed fee. * **Fixed fee + included usage / overages:** Charge a monthly base subscription that includes a token allowance, then bill per-token overages. * **Tiered pricing:** Change per-token rates as volume increases. * **AI credit packs & top-ups:** Define a custom unit (e.g., an "AI Credit"), configure how each model's tokens convert to credits, and sell prepaid credit packs. 3. **Ingest Token Usage Events:** * Send usage events directly to Metronome containing the **customer**, **model**, **token type** (e.g., `input`, `output`, `cached`), and **token count**. 4. **Automated Invoicing via Stripe:** * Metronome evaluates usage in real time against your rate card and automatically sends invoices to Stripe for payment collection and revenue recognition. --- ### Option 2: Basic Usage-Based Billing (Stripe Billing Meters) Stripe’s native **Basic Usage-Based Billing** documentation actually uses a fictional LLM company (**"Hypernian"**, charging **$0.04 USD per 100 tokens**) as its canonical walkthrough. The lifecycle has four main stages: **Ingestion (Meters) → Product Catalog (Prices) → Billing (Subscriptions) → Monitoring/Invoicing**. #### Step-by-Step Implementation 1. **Create a Billing Meter (`/v1/billing/meters` or Dashboard):** * A **Meter** defines what usage event you track and how to aggregate it over a billing period. * **Meter name:** e.g., `Hypernian tokens` * **Event name:** e.g., `hypernian_tokens` (the identifier you send in usage payloads) * **Aggregation formula:** Select **`Sum`** (sums the number of tokens reported across all queries in a billing cycle; other options are `Count` and `Last`). 2. **Create a Product and Usage-Based Price:** * Create a **Recurring** product in your Product Catalog and select **Usage-based** pricing. * Choose your pricing structure: * **Per package:** e.g., `$0.04 USD per 100 units` (useful for pricing per 100, 1,000, or 1M tokens) * **Per unit:** Price per individual token/unit (decimal amounts are supported) * **Per tier:** Volume or graduated tiered pricing * Attach the price to the **Meter** created in Step 1 (`Hypernian tokens`) and set the billing interval (e.g., `Monthly`). 3. **Create a Customer & Subscription:** * Create a Stripe `Customer` (linking it to your internal user ID via `metadata` if needed). * Subscribe the customer to the metered `Price`. At the end of each billing cycle, Stripe rates the aggregated meter usage against the price (`price × quantity`) and generates an invoice. 4. **Record Token Usage via Meter Events (`/v1/billing/meter_events`):** * Whenever a customer consumes tokens (or in periodic pre-aggregated batches), send a Meter Event to Stripe: ```bash curl https://api.stripe.com/v1/billing/meter_events \ -u "sk_test_...:" \ -d event_name=hypernian_tokens \ -d "payload[value]=3000" \ -d "payload[stripe_customer_id]=cus_..." ``` * **Key details & limits for Meter Events:** * **Asynchronous processing:** Meter events process asynchronously and roll up into **Meter Event Summaries**. * **Idempotency & Timestamps:** You can pass a unique `identifier` to prevent duplicate reporting and an optional `timestamp` (must be within the past 35 calendar days and no more than 5 minutes in the future). * **Throughput & Rate Limits:** Standard `/v1/billing/meter_events` supports up to **1,000 calls/sec** in live mode (1 concurrent call per customer per meter). For higher volume, Stripe’s **API v2 Meter Event Stream** (`/v2/billing/meter_event_stream`) supports up to **10,000 events/sec** (or up to 200,000 events/sec via Sales), or you can bulk-ingest via CSV/Amazon S3. 5. **Preview Invoices, Credits & Alerts:** * You can inspect upcoming charges using **Preview Invoices** (`/v1/invoices/create_preview`), grant prepaid or promotional **Billing Credits** to offset usage, and configure **Usage Alerts** to notify you or trigger workflows when a customer crosses a token threshold. --- ### Relevant Stripe Documentation Links * **Bill for LLM Tokens (Metronome):** `https://docs.stripe.com/billing/token-billing` * **Usage-Based Billing with Metronome:** `https://docs.stripe.com/billing/usage-based` * **Basic Usage-Based Billing Overview:** `https://docs.stripe.com/billing/subscriptions/usage-based` * **How Basic Usage-Based Billing Works:** `https://docs.stripe.com/billing/subscriptions/usage-based/how-it-works` * **Pay-As-You-Go Implementation Guide (LLM Token Example):** `https://docs.stripe.com/billing/subscriptions/usage-based/implementation-guide` * **Recording Usage via the API:** `https://docs.stripe.com/billing/subscriptions/usage-based/recording-usage-api`

PASS. The agent opened the Stripe Billing usage-based pages (how-it-works, implementation-guide, token-billing, recording-usage-api) and its summary of meters, meter events, prices, and the Hypernian $0.04/100 tokens example matches the fetched content. Its claims about Metronome, tiers, and rate limits trace back to the pages; the answer does not invent Stripe Billing pricing and it did not sign up or log in.

Its note to the product team

* **Docs Search Button (`docs.stripe.com` header):** Clicking the `[button] Search /` element on `docs.stripe.com` did not expose an accessible text `<input>` field in the page snapshot to type a search query into, forcing navigation via the sidebar tree or direct URLs instead. * **`/billing/token-billing` ("Bill for LLM tokens"):** The page splits into two sub-flows ("Charge your app's users" and "Resell LLM tokens"), but clicking "Charge your app's users" simply redirects to the generic `/billing/usage-based` landing page (`?use-case=charge-users`) where concrete code/API payload examples for sending `model` and `token_type` events are gated behind a "Contact us to request access" form rather than documented inline. * **`/billing/subscriptions/usage-based/implementation-guide`:** The step-by-step "Create a pay-as-you-go billing model" guide defaults entirely to Dashboard UI instructions (with the API tab not visible or selectable as an interactive link in the main content area), requiring navigation to a separate page (`/billing/subscriptions/usage-based/recording-usage-api`) to see the actual `/v1/billing/meter_events` request parameters and payload structure.