Stripe AI

https://github.com/stripe/ai
Admin
Stripe AI provides SDKs and tools for integrating Stripe APIs with LLMs and agent frameworks,...
Tokens:111,904
Snippets:1,832
Trust Score:9.7
License:MIT
Update:2 months ago
Show doc for...
Tokens:
Raw
### Install Dependencies and Start Server

Source: https://github.com/stripe/ai/blob/main/benchmarks/subscription-gym/README.md

Navigate to the server directory, install Ruby dependencies using Bundler, and start the Sinatra server.

```bash
cd environment/server
bundle install
bundle exec ruby server.rb
```

--------------------------------

### Install Dependencies and Start Server

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-ruby/README.md

Install Ruby dependencies using Bundler and start the local server. Verify it's running by curling the /config endpoint.

```bash
cd environment/$PROBLEM/server
bundle install
bundle exec ruby server.rb
```

```bash
curl http://localhost:4242/config
```

--------------------------------

### Setup Environment Variables

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/meter/examples/README.md

Copy the example environment file and add your Stripe and provider-specific API keys.

```bash
cd examples
cp .env.example .env
```

```bash
# Required for all examples
STRIPE_API_KEY=sk_test_...
STRIPE_CUSTOMER_ID=cus_...

# Provider-specific (only required for the provider you want to test)
OPENAI_API_KEY=sk-...
ANTHROPIC_API_KEY=sk-ant-...
GOOGLE_GENERATIVE_AI_API_KEY=...
```

--------------------------------

### Setup Development Environment

Source: https://github.com/stripe/ai/blob/main/tools/python/README.md

Set up a Python virtual environment and install development dependencies.

```sh
uv venv --python 3.11
source .venv/bin/activate
uv pip install -r requirements.txt
```

--------------------------------

### Install Dependencies

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/meter/examples/README.md

Navigate to the ai-sdk-billing-wrapper directory and install its dependencies.

```bash
cd llm/ai-sdk-billing-wrapper
npm install
```

--------------------------------

### Start Local Server

Source: https://github.com/stripe/ai/blob/main/benchmarks/galtee-invoicing/README.md

Navigate to the server directory and start the Node.js server.

```bash
cd environment/server
npm start
```

--------------------------------

### Run OpenAI Examples

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/provider/examples/README.md

Execute the OpenAI examples using ts-node from the examples directory.

```bash
cd examples
npx ts-node openai.ts
```

--------------------------------

### Run the Sinatra Application

Source: https://github.com/stripe/ai/blob/main/benchmarks/checkout-gym/environment/server/README.md

Execute this command to start the Sinatra server. Ensure you have a configured .env file and all dependencies installed.

```bash
ruby server.rb
```

--------------------------------

### Install Server Dependencies

Source: https://github.com/stripe/ai/blob/main/benchmarks/galtee-invoicing/README.md

Navigate to the server directory and install Node.js dependencies using npm.

```bash
cd environment/server
npm install
```

--------------------------------

### Start the Server

Source: https://github.com/stripe/ai/blob/main/benchmarks/card-element-to-checkout/README.md

Starts the Flask development server. Ensure the .env file is configured with your Stripe API keys before running.

```bash
cd environment/server

# Create .env from the example, then fill in your Stripe keys
cp .env.example .env
# Edit .env and set STRIPE_SECRET_KEY and STRIPE_PUBLISHABLE_KEY

python main.py
```

--------------------------------

### Start Development Server

Source: https://github.com/stripe/ai/blob/main/benchmarks/saas-starter-embedded-checkout/environment/README.md

Run the Next.js development server to view the application locally.

```bash
pnpm dev
```

--------------------------------

### Install Server Dependencies

Source: https://github.com/stripe/ai/blob/main/benchmarks/card-element-to-checkout/README.md

Installs the necessary Python dependencies for the server environment.

```bash
cd environment/server
pip install -e .
```

--------------------------------

### Install uv Package Manager

Source: https://github.com/stripe/ai/blob/main/tools/python/examples/openai/customer_support/README.md

Installs the uv package manager using a curl script. Ensure uv is installed before proceeding with project setup.

```bash
curl -LsSf https://astral.sh/uv/install.sh | sh
```

--------------------------------

### Build and Run the Server

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/README.md

Navigate to the problem's server directory, restore dependencies, and run the .NET application. The server will start on http://localhost:4242.

```bash
cd environment/$PROBLEM/server
dotnet restore
dotnet run
```

--------------------------------

### Run the Server

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/environment/charges-on-payment-intent/server/README.md

Execute the .NET application to start the server. The server will be accessible at `http://localhost:4242`.

```bash
dotnet run
```

--------------------------------

### Start MongoDB Service

Source: https://github.com/stripe/ai/blob/main/benchmarks/furever/environment/README.md

Start the MongoDB Community Edition service using Homebrew. This command assumes MongoDB has been installed via Homebrew.

```bash
brew services start mongodb-community@7.0
```

--------------------------------

### Install Stripe Agent Toolkit

Source: https://github.com/stripe/ai/blob/main/tools/python/README.md

Install the library using uv pip. Ensure you have Python 3.11+.

```sh
uv pip install stripe-agent-toolkit
```

--------------------------------

### Verify Skill Installation

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/stripe-projects/SKILL.md

Confirm that the stripe-projects-cli skill has been correctly installed locally.

```bash
test -f .claude/skills/stripe-projects-cli/SKILL.md && echo "OK" || echo "MISSING"
```

--------------------------------

### Install Server Dependencies - Bash

Source: https://github.com/stripe/ai/blob/main/benchmarks/checkout-gym/README.md

Navigate to the server directory and install the necessary Ruby dependencies using Bundler.

```bash
cd environment/server
bundle install
```

--------------------------------

### Start Development Server

Source: https://github.com/stripe/ai/blob/main/benchmarks/furever/environment/README.md

Run this command in your terminal to start the application's development server. Access the app via http://localhost:{process.env.PORT}.

```bash
npm run dev &
```

--------------------------------

### Install Dependencies with Bundler

Source: https://github.com/stripe/ai/blob/main/benchmarks/checkout-gym/environment/server/README.md

Use this command to install all project dependencies listed in the Gemfile.

```bash
bundle install
```

--------------------------------

### Run Google Gemini Example

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/meter/examples/README.md

Execute the Google Gemini example using ts-node to test Gemini models with the billing wrapper.

```bash
npx ts-node google.ts
```

--------------------------------

### Environment Variables Setup

Source: https://github.com/stripe/ai/blob/main/tools/python/examples/openai/file_search/README.md

Configure your API keys and vector store ID by copying the template and populating the values.

```bash
OPENAI_API_KEY=your_openai_api_key
OPENAI_VECTOR_STORE_ID=your_openai_vector_store_id
STRIPE_SECRET_KEY=your_stripe_secret_key
```

--------------------------------

### Copy Environment Example

Source: https://github.com/stripe/ai/blob/main/benchmarks/furever/README.md

Copies the example environment file to be used for configuration. Remember to replace placeholder keys with your actual Stripe test keys.

```bash
cp environment/.env.example environment/.env
```

--------------------------------

### Clone and Install Project Dependencies

Source: https://github.com/stripe/ai/blob/main/benchmarks/saas-starter-embedded-checkout/environment/README.md

Clone the repository and install project dependencies using pnpm.

```bash
git clone https://github.com/nextjs/saas-starter
cd saas-starter
pnpm install
```

--------------------------------

### Run Solution Server

Source: https://github.com/stripe/ai/blob/main/benchmarks/card-element-to-checkout/README.md

Builds the Docker image, initializes Stripe products, and starts the solution server on localhost:5000. Use this to interact with the storefront in a browser and manually complete checkout for grading.

```bash
STRIPE_SECRET_KEY=sk_test_xxx STRIPE_PUBLISHABLE_KEY=pk_test_xxx ./run_server.sh
```

--------------------------------

### Run OpenAI Example

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/meter/examples/README.md

Execute the OpenAI example using ts-node to test various OpenAI features with the billing wrapper.

```bash
npx ts-node openai.ts
```

--------------------------------

### Example Response Schema for GET /products

Source: https://github.com/stripe/ai/blob/main/benchmarks/galtee-invoicing/environment/PROBLEM.md

This JSON structure represents the expected response for the GET /products API endpoint, detailing hike products, their Stripe IDs, pricing in multiple currencies, and names.

```json
[
  {
    "id": "fr_hike",
    "stripe_product_id": "prod_testproduct",
    "prices": {
      "eur": 1000,
      "usd": 3000
    },
    "name": "French Alps hike"
  },
  ... // additional products
]
```

--------------------------------

### Install Dependencies

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/provider/examples/README.md

Install the necessary dependencies for the Stripe AI SDK Provider from the ai-sdk-provider directory.

```bash
cd llm/ai-sdk-provider
npm install
```

--------------------------------

### Install Project Dependencies

Source: https://github.com/stripe/ai/blob/main/tools/python/examples/openai/customer_support/README.md

Installs all project dependencies defined in the environment using uv. This command should be run after activating the virtual environment.

```bash
uv sync
```

--------------------------------

### Install Stripe Projects Plugin

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/stripe-projects/SKILL.md

Install the required projects plugin for the Stripe CLI.

```bash
stripe plugin install projects
```

--------------------------------

### Install @stripe/token-meter for Native SDKs

Source: https://github.com/stripe/ai/blob/main/llm/README.md

Use this command to install the @stripe/token-meter package for direct integration with native AI SDKs.

```bash
npm install @stripe/token-meter
```

--------------------------------

### Configure Environment Variables

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/provider/examples/README.md

Copy the example environment file and add your Stripe API key and Customer ID.

```bash
cd examples
cp .env.example .env
```

--------------------------------

### Install Stripe.js via npm

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/upgrade-stripe/SKILL.md

Install the Stripe.js library using the npm package manager.

```bash
npm install @stripe/stripe-js
```

--------------------------------

### Check Tool Installation

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/stripe-directory/SKILL.md

Verify if a specific payment tool is installed on the system before proceeding with a purchase.

```bash
which <tool> 2>/dev/null
```

--------------------------------

### Copy .env.example for Server Configuration

Source: https://github.com/stripe/ai/blob/main/benchmarks/subscription-gym/README.md

Copy the example .env file to the server directory and edit it with your Stripe test keys.

```bash
cp .env.example environment/server/.env
```

--------------------------------

### Docker Setup - Using .env file

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/README.md

Create a .env file in the root directory with Stripe API keys and then execute the run_solution.sh script to test the Docker workflow.

```bash
cat > .env << 'EOF'
STRIPE_SECRET_KEY=sk_test_...
STRIPE_PUBLISHABLE_KEY=pk_test_...
EOF
```

```bash
./run_solution.sh
```

--------------------------------

### Initialize Stripe Products

Source: https://github.com/stripe/ai/blob/main/benchmarks/card-element-to-checkout/README.md

Initializes Stripe products and prices. This is a one-time setup step performed from the project root directory.

```python
STRIPE_SECRET_KEY=sk_test_xxx python environment/init_products.py
```

--------------------------------

### Install Link CLI

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/stripe-directory/SKILL.md

Command to install the Stripe Link CLI globally via npm.

```bash
npm i -g @stripe/link-cli
```

--------------------------------

### Install @stripe/ai-sdk for Vercel AI SDK

Source: https://github.com/stripe/ai/blob/main/llm/README.md

Use this command to install the @stripe/ai-sdk package when integrating with the Vercel AI SDK.

```bash
npm install @stripe/ai-sdk
```

--------------------------------

### Install Dependencies with npm

Source: https://github.com/stripe/ai/blob/main/benchmarks/furever/environment/README.md

Install project dependencies using npm. This command should be run in the project's root directory.

```bash
npm install
```

--------------------------------

### Run Main Application

Source: https://github.com/stripe/ai/blob/main/tools/python/examples/crewai/README.md

Execute the main application script using Python. Ensure all dependencies are installed and the .env file is correctly configured.

```python
python main.py
```

--------------------------------

### Configure Stripe API Keys

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-ruby/README.md

Set up your Stripe API keys by copying the example .env file and filling in your secret and publishable keys.

```bash
# Pick a problem to work on
PROBLEM=charges-on-payment-intent  # or invoice-partial-payments, subscription-billing-migration

# Create .env from the example
cp environment/$PROBLEM/server/.env.example environment/$PROBLEM/server/.env
```

```bash
STRIPE_SECRET_KEY=sk_test_...
STRIPE_PUBLISHABLE_KEY=pk_test_...
```

--------------------------------

### Install Stripe plugin for Grok Build

Source: https://github.com/stripe/ai/blob/main/README.md

Use this command to install the Stripe plugin for Grok Build with trust permissions.

```bash
grok plugin install stripe --trust
```

--------------------------------

### Install Grader Dependencies

Source: https://github.com/stripe/ai/blob/main/benchmarks/subscription-gym/README.md

Install the necessary Ruby dependencies for the grader using Bundler. This only needs to be done once.

```bash
cd grader
bundle install
```

--------------------------------

### Verify Stripe CLI Installation

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/stripe-projects/SKILL.md

Check if the Stripe CLI is installed and verify its version.

```bash
which stripe && stripe --version
```

--------------------------------

### Run Anthropic Claude Example

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/meter/examples/README.md

Execute the Anthropic Claude example using ts-node to test Claude models with the billing wrapper.

```bash
npx ts-node anthropic.ts
```

--------------------------------

### Environment Variables Setup

Source: https://github.com/stripe/ai/blob/main/benchmarks/furever/README.md

Configuration of Stripe API keys in the environment file. Ensure you use your real Stripe test keys.

```bash
STRIPE_SECRET_KEY="sk_test_..."
STRIPE_PUBLIC_KEY="pk_test_..."
```

--------------------------------

### Configure Environment Variables

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/environment/charges-on-payment-intent/server/README.md

Navigate to the server directory and copy the example environment file. Edit the `.env` file to include your Stripe test API keys.

```bash
cd environment/server
cp .env.example .env
# Edit .env and add your Stripe test API keys
```

--------------------------------

### Install Stripe plugin for Claude Code

Source: https://github.com/stripe/ai/blob/main/README.md

Use this command to install the official Stripe plugin for Claude Code.

```bash
claude plugin install stripe@claude-plugins-official
```

--------------------------------

### Basic Stripe AI SDK Setup

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/provider/examples/README.md

Initialize the Stripe AI SDK provider with your API key and an optional default customer ID. Then, create a model instance.

```typescript
import { createStripe } from '@stripe/ai-sdk/provider';
import { generateText } from 'ai';

const stripeLLM = createStripe({
  apiKey: process.env.STRIPE_API_KEY!,
  customerId: process.env.STRIPE_CUSTOMER_ID, // Optional default
});

const model = stripe('openai/gpt-5');
```

--------------------------------

### Create a connected account using Accounts v2

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/connect-recommend/references/account-types.md

Examples for creating marketplace and SaaS connected accounts using the v2 API.

```javascript
const account = await stripe.v2.core.accounts.create({
  contact_email: 'seller@example.com',
  display_name: 'Seller Name',
  dashboard: 'express',
  identity: { country: 'us', entity_type: 'individual' },
  configuration: {
    recipient: {
      capabilities: {
        stripe_balance: { stripe_transfers: { requested: true } },
      },
    },
  },
  defaults: {
    currency: 'usd',
    responsibilities: {
      fees_collector: 'application',
      losses_collector: 'application',
    },
  },
});
```

```javascript
const account = await stripe.v2.core.accounts.create({
  contact_email: 'merchant@example.com',
  display_name: 'Merchant Name',
  dashboard: 'full',
  identity: { country: 'us', entity_type: 'individual' },
  configuration: {
    merchant: {
      capabilities: {
        card_payments: { requested: true },
      },
    },
  },
  defaults: {
    currency: 'usd',
    responsibilities: {
      fees_collector: 'stripe',
      losses_collector: 'stripe',
    },
  },
});
```

--------------------------------

### Install Stripe plugin for Cursor

Source: https://github.com/stripe/ai/blob/main/README.md

Use this command within the Cursor environment to add the Stripe plugin.

```bash
/add-plugin stripe
```

--------------------------------

### Copy Environment File

Source: https://github.com/stripe/ai/blob/main/benchmarks/furever/environment/README.md

Copy the example environment file to .env to configure your Stripe API keys. This step can be skipped if a .env file already exists and is configured.

```bash
cp .env.example .env
```

--------------------------------

### Configure API Keys - Bash

Source: https://github.com/stripe/ai/blob/main/benchmarks/checkout-gym/README.md

Before generating new price IDs, copy the example environment file and edit it to include your Stripe test API keys.

```bash
cp environment/server/.env.example environment/server/.env
# Edit .env and add your Stripe test API keys
```

--------------------------------

### Stripe Meter Input Tokens Event Example

Source: https://github.com/stripe/ai/blob/main/llm/token-meter/README.md

An example of the JSON structure for a Stripe meter event reporting input token usage.

```json
{
  "event_name": "token-billing-tokens",
  "payload": {
    "stripe_customer_id": "cus_xxxxx",
    "value": "100",
    "model": "openai/gpt-4o-mini",
    "token_type": "input"
  }
}
```

--------------------------------

### Stripe Meter Output Tokens Event Example

Source: https://github.com/stripe/ai/blob/main/llm/token-meter/README.md

An example of the JSON structure for a Stripe meter event reporting output token usage.

```json
{
  "event_name": "token-billing-tokens",
  "payload": {
    "stripe_customer_id": "cus_xxxxx",
    "value": "50",
    "model": "openai/gpt-4o-mini",
    "token_type": "output"
  }
}
```

--------------------------------

### Get Configuration

Source: https://github.com/stripe/ai/blob/main/benchmarks/subscription-gym/environment/server/README.md

Returns the Stripe publishable key for client-side integration.

```APIDOC
## GET /config

### Description
Returns the Stripe publishable key for client-side integration.

### Method
GET

### Endpoint
/config

### Response
#### Success Response (200)
- **publishableKey** (string) - The Stripe publishable key.
```

--------------------------------

### Manually install agent skills

Source: https://github.com/stripe/ai/blob/main/README.md

Use this command to manually add Stripe skills to your project. Note that manual installations do not auto-update.

```bash
npx skills add https://docs.stripe.com
```

--------------------------------

### Docker Setup - Using environment variables

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/README.md

Alternatively, set the Stripe API keys as environment variables before running the run_solution.sh script for the Docker workflow.

```bash
export STRIPE_SECRET_KEY=sk_test_...
./run_solution.sh
```

--------------------------------

### Docker Setup with Environment Variables

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-ruby/README.md

Alternatively, set Stripe API keys as environment variables before running the Docker solution script.

```bash
STRIPE_SECRET_KEY=sk_test_... ./run_solution.sh
```

--------------------------------

### Initialize Toolkit with Constructor

Source: https://github.com/stripe/ai/blob/main/tools/typescript/MIGRATION.md

Use the `StripeAgentToolkit` constructor for separate instance creation and initialization, useful for dependency injection or delayed setup. Remember to call `initialize()` and `close()`.

```typescript
import StripeAgentToolkit from '@stripe/agent-toolkit/openai';

const toolkit = new StripeAgentToolkit({
  secretKey: 'rk_test_...',
  configuration: {},
});

// Later, when ready to use:
await toolkit.initialize();

const tools = toolkit.getTools();
// ... use tools ...

await toolkit.close(); // Clean up when done
```

--------------------------------

### Run TypeScript Example

Source: https://github.com/stripe/ai/blob/main/tools/typescript/examples/ai-sdk/README.md

Execute the main TypeScript file using ts-node. The --env flag indicates that environment variables should be loaded.

```bash
npx ts-node index.ts --env
```

--------------------------------

### Run the Grader

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/README.md

Copy the .env file to the grader directory, install Ruby dependencies, and execute the RSpec tests against the running server.

```bash
# Copy .env so the grader can access Stripe API keys
cp environment/$PROBLEM/server/.env grader/.env

# Install grader dependencies (first time only)
cd grader
bundle install

# Run the grader RSpec tests directly against the running server
bundle exec rspec $PROBLEM/grade.rb --format documentation
```

--------------------------------

### Get Products

Source: https://github.com/stripe/ai/blob/main/benchmarks/galtee-basic/environment/PROBLEM.md

Retrieves a list of all products offered by Galtee in JSON format.

```APIDOC
## GET /products

### Description
Retrieves a list of all products offered in JSON format according to the existing products table in the database.

### Method
GET

### Endpoint
/products
```

--------------------------------

### Configure Stripe API Keys

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/README.md

Set the PROBLEM environment variable to select the target problem and copy the example .env file. Then, edit the .env file with your Stripe test keys.

```bash
# Pick a problem to work on
export PROBLEM=charges-on-payment-intent  # or invoice-partial-payments, subscription-billing-migration

# Create .env from the example (or edit the existing .env)
cp environment/$PROBLEM/server/.env.example environment/$PROBLEM/server/.env
```

```bash
STRIPE_SECRET_KEY=sk_test_...
STRIPE_PUBLISHABLE_KEY=pk_test_...
```

--------------------------------

### Install Stripe plugin for Codex

Source: https://github.com/stripe/ai/blob/main/README.md

Use this command to add the curated Stripe plugin to your Codex project.

```bash
codex plugin add stripe@openai-curated
```

--------------------------------

### Configure Claude Desktop with Stripe MCP

Source: https://github.com/stripe/ai/blob/main/tools/modelcontextprotocol/README.md

Add the Stripe MCP server configuration to your claude_desktop_config.json. This example shows configuration using npx. Ensure STRIPE_KEY is replaced with your actual key.

```json
{
  "mcpServers": {
    "stripe": {
      "command": "npx",
      "args": ["-y", "@stripe/mcp", "--api-key=STRIPE_KEY"]
    }
  }
}
```

--------------------------------

### Set up Environment Variables

Source: https://github.com/stripe/ai/blob/main/benchmarks/saas-starter-embedded-checkout/environment/README.md

Use the provided script to create the .env file for local development.

```bash
pnpm db:setup
```

--------------------------------

### Run the REPL for Agent Testing

Source: https://github.com/stripe/ai/blob/main/tools/python/examples/openai/customer_support/README.md

Starts the Read-Eval-Print Loop (REPL) for interactively testing the support agent. This is useful for debugging and experimenting without sending emails.

```bash
python repl.py
```

--------------------------------

### Get Customer Bookings

Source: https://github.com/stripe/ai/blob/main/benchmarks/galtee-basic/environment/PROBLEM.md

Lists all purchases for a given customer, including their status.

```APIDOC
## GET /customer/:email/bookings

### Description
Lists all purchases for the given customer. This should include the data from the bookings table in addition to a `status` field. `status` should be ‘complete’ if fully paid, ‘refunded’ if refunded, ‘incomplete’ if incomplete.

### Method
GET

### Endpoint
/customer/:email/bookings

### Parameters
#### Path Parameters
- **email** (string) - Required - The email of the customer whose bookings are to be listed.
```

--------------------------------

### Install MCP Package Dependency

Source: https://github.com/stripe/ai/blob/main/tools/python/MIGRATION.md

The `mcp` package is now a required dependency for the Stripe Agent Toolkit. Install or upgrade using pip.

```bash
pip install stripe-agent-toolkit>=0.7.0
```

--------------------------------

### List All Charges

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/environment/charges-on-payment-intent/server/README.md

Retrieve a list of all charges associated with a given `payment_intent_id` by sending a GET request to the `/payment-details` endpoint.

```bash
curl "http://localhost:4242/payment-details?payment_intent_id=pi_..."
```

--------------------------------

### Initialize Stripe Agent Toolkit for Model Context Protocol

Source: https://github.com/stripe/ai/blob/main/tools/typescript/README.md

Instantiate the StripeAgentToolkit for use with the Model Context Protocol (MCP). This setup is for running an MCP server.

```typescript
import {StripeAgentToolkit} from '@stripe/agent-toolkit/modelcontextprotocol';
import {StdioServerTransport} from '@modelcontextprotocol/sdk/server/stdio.js';

const server = new StripeAgentToolkit({
  secretKey: process.env.STRIPE_SECRET_KEY!,
  configuration: {},
});

async function main() {
  const transport = new StdioServerTransport();
  await server.connect(transport);
  console.error('Stripe MCP Server running on stdio');
}

main().catch((error) => {
  console.error('Fatal error in main():', error);
  process.exit(1);
});
```

--------------------------------

### Save Payment Method for Future Use

Source: https://github.com/stripe/ai/blob/main/providers/kiro/get-started-kiro.md

This snippet shows how to set up a customer to save their payment method for future charges using Setup Intents. It also includes creating a payment intent with a saved method.

```javascript
// Step 1: Create Setup Intent
const setupIntent = createSetupIntent({
  customer: "cus_xxx",
  payment_method_types: ["card"],
});

// Step 2: Collect payment method on frontend with Setup Intent
// Step 3: Handle webhook for setup_intent.succeeded
// Step 4: Use saved payment method for future charges
const paymentIntent = createPaymentIntent({
  amount: 2999,
  currency: "usd",
  customer: "cus_xxx",
  payment_method: "pm_xxx",
  off_session: true,
  confirm: true,
});
```

--------------------------------

### Get Latest Charge Details

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/environment/charges-on-payment-intent/server/README.md

Fetch the details of the latest charge for a specific `payment_intent_id` via a GET request to the `/latest-charge-details` endpoint.

```bash
curl "http://localhost:4242/latest-charge-details?payment_intent_id=pi_..."
```

--------------------------------

### Create Subscription Request Body

Source: https://github.com/stripe/ai/blob/main/benchmarks/subscription-gym/environment/server/README.md

Example JSON payload for creating a new subscription via the POST /create-subscription endpoint.

```json
{
  "challenge": "basic_subscription"
}
```

--------------------------------

### Create Stripe Agent Toolkit with CrewAI

Source: https://github.com/stripe/ai/blob/main/tools/python/MIGRATION.md

Use the `create_stripe_agent_toolkit` factory function with `await` and ensure `await toolkit.close()` is called in a `finally` block for proper cleanup. This example demonstrates creating a Stripe Agent for product creation using CrewAI.

```python
import asyncio
from crewai import Agent, Task, Crew
from stripe_agent_toolkit.crewai.toolkit import create_stripe_agent_toolkit

async def main():
    toolkit = await create_stripe_agent_toolkit(secret_key="rk_test_...")

    try:
        agent = Agent(
            role="Stripe Agent",
            goal="Create Stripe products",
            tools=toolkit.get_tools(),
        )
        task = Task(description="Create a product", agent=agent)
        crew = Crew(agents=[agent], tasks=[task])
        crew.kickoff()
    finally:
        await toolkit.close()

asyncio.run(main())
```

--------------------------------

### Get Products

Source: https://github.com/stripe/ai/blob/main/benchmarks/galtee-invoicing/environment/PROBLEM.md

Retrieves a list of all available hike products offered by Galtee, including their Stripe product IDs and pricing in various currencies.

```APIDOC
## GET /products

### Description
Retrieves a list of products offered in JSON format, including their associated Stripe product IDs and pricing details per currency.

### Method
GET

### Endpoint
/products

### Response
#### Success Response (200)
- **id** (string) - The unique identifier for the product (e.g., "fr_hike").
- **stripe_product_id** (string) - The identifier for the product in the Stripe API.
- **prices** (object) - An object containing currency codes as keys and the price in cents as values (e.g., {"eur": 1000, "usd": 3000}).
- **name** (string) - The display name of the product (e.g., "French Alps hike").

#### Response Example
```json
[
  {
    "id": "fr_hike",
    "stripe_product_id": "prod_testproduct",
    "prices": {
      "eur": 1000,
      "usd": 3000
    },
    "name": "French Alps hike"
  }
]
```
```

--------------------------------

### Get Subscription Status

Source: https://github.com/stripe/ai/blob/main/benchmarks/subscription-gym/environment/server/README.md

Retrieves the status of a subscription using its ID.

```APIDOC
## GET /subscription-status

### Description
Retrieves the status of a subscription.

### Method
GET

### Endpoint
/subscription-status

### Parameters
#### Query Parameters
- **subscription_id** (string) - Required - The ID of the subscription to retrieve the status for.

### Response
#### Success Response (200)
- **status** (string) - The current status of the subscription.
```

--------------------------------

### Multi-turn conversations with meteredModel

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/meter/README.md

Use `meteredModel` to track token usage in multi-turn conversations. This example demonstrates with the Google Gemini provider.

```typescript
import { meteredModel } from '@stripe/ai-sdk/meter';
import { google } from '@ai-sdk/google';
import { generateText } from 'ai';

const model = meteredModel(
  google('gemini-2.5-flash'),
  process.env.STRIPE_API_KEY,
  'cus_xxxxx'
);

const result = await generateText({
  model,
  messages: [
    { role: 'user', content: 'What is the capital of France?' },
    { role: 'assistant', content: 'The capital of France is Paris.' },
    { role: 'user', content: 'What is its population?' },
  ],
});
```

--------------------------------

### Stream Text with Stripe Provider

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/provider/README.md

Stream text output from a specified model using the Stripe provider. This example uses Google Gemini and iterates over the text stream.

```typescript
import { createStripe } from '@stripe/ai-sdk/provider';
import { streamText } from 'ai';

const stripeLLM = createStripe({
  apiKey: process.env.STRIPE_API_KEY,
  customerId: 'cus_xxxxx',
});

const result = streamText({
  model: stripe('google/gemini-2.5-flash'),
  prompt: 'Write a short story about AI.',
});

for await (const chunk of result.textStream) {
  process.stdout.write(chunk);
}
```

--------------------------------

### Streaming responses with meteredModel

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/meter/README.md

Enable automatic usage tracking for streaming responses by wrapping the AI SDK model with `meteredModel`. This example uses the Anthropic provider.

```typescript
import { meteredModel } from '@stripe/ai-sdk/meter';
import { anthropic } from '@ai-sdk/anthropic';
import { streamText } from 'ai';

const model = meteredModel(
  anthropic('claude-3-5-sonnet-20241022'),
  process.env.STRIPE_API_KEY,
  'cus_xxxxx'
);

const result = streamText({
  model,
  prompt: 'Write a short story about AI.',
});

for await (const chunk of result.textStream) {
  process.stdout.write(chunk);
}
```

--------------------------------

### Quick Start: Track OpenAI Chat Completions

Source: https://github.com/stripe/ai/blob/main/llm/token-meter/README.md

Integrate Stripe Token Meter with OpenAI for non-streaming chat completions. Ensure your OpenAI API key and Stripe API key are set as environment variables. The `meter.trackUsage` function sends billing events asynchronously.

```typescript
import OpenAI from 'openai';
import { createTokenMeter } from '@stripe/token-meter';

const openai = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });
const meter = createTokenMeter(process.env.STRIPE_API_KEY);

// Non-streaming
const response = await openai.chat.completions.create({
  model: 'gpt-4o-mini',
  messages: [{ role: 'user', content: 'Hello!' }],
});

meter.trackUsage(response, 'cus_xxxxx');
```

--------------------------------

### Initialize Stripe Project

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/stripe-projects/SKILL.md

Check project status, run preflight checks, and initialize a new project.

```bash
stripe projects status --json
```

```bash
stripe projects init --preflight --json
```

```bash
stripe projects init --accept-tos --yes
```

--------------------------------

### Initialize Stripe Agent Toolkit (Async)

Source: https://github.com/stripe/ai/blob/main/tools/python/MIGRATION.md

Migrate from synchronous initialization to asynchronous initialization using `create_stripe_agent_toolkit`. Ensure to await the initialization and close the toolkit when done.

```python
# Before (v0.6.x)
from stripe_agent_toolkit.openai.toolkit import StripeAgentToolkit

toolkit = StripeAgentToolkit(secret_key="sk_test_...", configuration={...})
tools = toolkit.tools

# After (v0.7.0+)
from stripe_agent_toolkit.openai.toolkit import create_stripe_agent_toolkit

toolkit = await create_stripe_agent_toolkit(secret_key="rk_test_...")
tools = toolkit.get_tools()
await toolkit.close()  # Clean up when done
```

--------------------------------

### Pay Invoice After Attaching Payments

Source: https://github.com/stripe/ai/blob/main/benchmarks/saas-starter-partial-payments/solution/SOLUTION_NOTES.md

Once all partial payments have been successfully attached to the invoice, call `stripe.invoices.pay` to finalize the invoice and complete the subscription setup.

```typescript
// After both payments are attached, pay the invoice
await stripe.invoices.pay(invoiceId);
```

--------------------------------

### Project Structure Overview

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/README.md

This diagram illustrates the directory structure for the SDK upgrade project, highlighting the 'environment', 'grader', and 'solution' folders.

```bash
sdk-upgrades-dotnet/
├── environment/          # Given to AI (old SDK code)
│   ├── charges-on-payment-intent/
│   ├── invoice-partial-payments/
│   └── subscription-billing-migration/
├── grader/               # Hidden (RSpec tests)
├── solution/             # Reference (upgraded SDK code)
├── Dockerfile
├── run_inside_docker.sh
└── run_solution.sh
```

--------------------------------

### Build and Run Java Server

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-java/README.md

Compile and run the Java server locally. Ensure you navigate to the server directory first.

```bash
cd environment/$PROBLEM/server
mvn install -DskipTests
mvn exec:java -Dexec.mainClass="com.stripe.sample.Server"
```

--------------------------------

### Quick Start: Track Google Gemini Content Generation

Source: https://github.com/stripe/ai/blob/main/llm/token-meter/README.md

Integrate Stripe Token Meter with Google Gemini for non-streaming content generation. Ensure your Google API key and Stripe API key are set as environment variables. The `meter.trackUsage` function is used to send billing events.

```typescript
import { GoogleGenerativeAI } from '@google/generative-ai';
import { createTokenMeter } from '@stripe/token-meter';

const genAI = new GoogleGenerativeAI(process.env.GOOGLE_API_KEY);
const meter = createTokenMeter(process.env.STRIPE_API_KEY);

const model = genAI.getGenerativeModel({ model: 'gemini-2.5-flash' });
const response = await model.generateContent('Hello!');

meter.trackUsage(response.response, 'cus_xxxxx');
```

--------------------------------

### Track Google Gemini Streaming Responses

Source: https://github.com/stripe/ai/blob/main/llm/token-meter/README.md

Use the token meter with Google Gemini's streaming API to track token consumption. This example shows how to wrap the stream for usage reporting.

```typescript
import { GoogleGenerativeAI } from '@google/generative-ai';
import { createTokenMeter } from '@stripe/token-meter';

const genAI = new GoogleGenerativeAI(process.env.GOOGLE_API_KEY);
const meter = createTokenMeter(process.env.STRIPE_API_KEY);

const model = genAI.getGenerativeModel({ model: 'gemini-2.5-flash' });
const result = await model.generateContentStream('Count to 5');

const meteredStream = meter.trackUsageStreamGemini(
  result,
  'cus_xxxxx',
  'gemini-2.5-flash'
);

for await (const chunk of meteredStream.stream) {
  process.stdout.write(chunk.text());
}
```

--------------------------------

### Track OpenAI Embeddings API Usage

Source: https://github.com/stripe/ai/blob/main/llm/token-meter/README.md

Use the token meter to track usage when calling the OpenAI Embeddings API. This example demonstrates tracking the response from an embeddings request.

```typescript
import OpenAI from 'openai';
import { createTokenMeter } from '@stripe/token-meter';

const openai = new OpenAI({ apiKey: process.env.OPENAI_API_KEY });
const meter = createTokenMeter(process.env.STRIPE_API_KEY);

const response = await openai.embeddings.create({
  model: 'text-embedding-3-small',
  input: 'Hello, world!',
});

meter.trackUsage(response, 'cus_xxxxx');
console.log('Embedding dimensions:', response.data[0].embedding.length);
```

--------------------------------

### Onboarding Method Selection Logic

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/connect-recommend/references/decision-matrix.md

Decision tree for choosing between embedded components and custom API-based onboarding flows.

```text
How much control over onboarding UX?
├── "Stripe handles everything" → Embedded components (recommended default)
├── "Some customization" → Embedded components with [appearance options API](/connect/embedded-appearance-options)
└── "Fully custom" → API-based — NOT RECOMMENDED for platforms integrating without dedicated Stripe guidance.
    Requires building custom remediation flows. Direct to [Stripe sales](https://stripe.com/contact/sales).
```

--------------------------------

### Quick Start: Track Anthropic Messages

Source: https://github.com/stripe/ai/blob/main/llm/token-meter/README.md

Integrate Stripe Token Meter with Anthropic for non-streaming messages. Ensure your Anthropic API key and Stripe API key are set as environment variables. The `meter.trackUsage` function handles billing event submission.

```typescript
import Anthropic from '@anthropic-ai/sdk';
import { createTokenMeter } from '@stripe/token-meter';

const anthropic = new Anthropic({ apiKey: process.env.ANTHROPIC_API_KEY });
const meter = createTokenMeter(process.env.STRIPE_API_KEY);

// Non-streaming
const response = await anthropic.messages.create({
  model: 'claude-3-5-sonnet-20241022',
  max_tokens: 1024,
  messages: [{ role: 'user', content: 'Hello!' }],
});

meter.trackUsage(response, 'cus_xxxxx');
```

--------------------------------

### Initialize Toolkit with Factory Function

Source: https://github.com/stripe/ai/blob/main/tools/typescript/MIGRATION.md

Use the `createStripeAgentToolkit` factory function for a recommended, one-step initialization process. Ensure to import from the correct SDK path.

```typescript
import {createStripeAgentToolkit} from '@stripe/agent-toolkit/openai';
// Also available: /ai-sdk, /langchain, /modelcontextprotocol

const toolkit = await createStripeAgentToolkit({
  secretKey: 'rk_test_...',
  configuration: {},
});

const tools = toolkit.getTools();
// ... use tools ...

await toolkit.close(); // Clean up when done
```

--------------------------------

### Checklist Update

Source: https://github.com/stripe/ai/blob/main/providers/codex/plugin/skills/connect-recommend/SKILL.md

Markdown representation of the completed onboarding steps.

```markdown
  [x] Learn about your business
  [x] Scan your project
  [x] Recommend configuration + charge pattern
  [x] Produce recommendation plan
```

--------------------------------

### Read a documentation page

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/stripe-docs/SKILL.md

Fetches and displays a specific documentation page by its web path.

```bash
stripe docs /payments
```

--------------------------------

### Initialize Toolkit using Constructor and initialize()

Source: https://github.com/stripe/ai/blob/main/tools/python/MIGRATION.md

Instantiate `StripeAgentToolkit` separately and call `initialize()` later for delayed initialization or dependency injection. Ensure to close the toolkit.

```python
from stripe_agent_toolkit.openai.toolkit import StripeAgentToolkit

toolkit = StripeAgentToolkit(secret_key="rk_test_...")

# Later, when ready to use:
async def main():
    await toolkit.initialize()
    try:
        tools = toolkit.get_tools()
        # ... use tools ...
    finally:
        await toolkit.close()
```

--------------------------------

### Initialize Stripe Agent Toolkit (Async)

Source: https://github.com/stripe/ai/blob/main/tools/typescript/MIGRATION.md

Before v0.9.0, initialization was synchronous. After v0.9.0, it requires `await` and includes a `close()` method for cleanup.

```typescript
// Before (v0.8.x)
const toolkit = new StripeAgentToolkit({secretKey, configuration});
const tools = toolkit.getTools();

// After (v0.9.0+)
const toolkit = await createStripeAgentToolkit({secretKey, configuration});
const tools = toolkit.getTools();
await toolkit.close(); // Clean up when done
```

--------------------------------

### Create Promotion Code via Stripe CLI

Source: https://github.com/stripe/ai/blob/main/benchmarks/saas-starter-embedded-checkout/solution/SOLUTION_NOTES.md

Use the Stripe CLI to create a coupon with a percentage discount and then associate it with a promotion code.

```bash
# Via Stripe CLI:
stripe coupons create --percent-off=20 --duration=once

# Then create promo code:
stripe promotion_codes create --coupon=<coupon_id> --code=TAKE20
```

--------------------------------

### Create and Activate Virtual Environment

Source: https://github.com/stripe/ai/blob/main/tools/python/examples/openai/customer_support/README.md

Creates a virtual environment using uv and activates it. This isolates project dependencies. Use the appropriate command for your operating system.

```bash
uv venv
source .venv/bin/activate  # On Unix/macOS
# or
.venv\Scripts\activate  # On Windows
```

--------------------------------

### Initialize Toolkit using Factory Function

Source: https://github.com/stripe/ai/blob/main/tools/python/MIGRATION.md

Use the `create_stripe_agent_toolkit` factory function for a simple, one-step toolkit creation and initialization. Remember to close the toolkit when finished.

```python
from stripe_agent_toolkit.openai.toolkit import create_stripe_agent_toolkit
# Also available: .langchain.toolkit, .crewai.toolkit, .strands.toolkit

async def main():
    toolkit = await create_stripe_agent_toolkit(secret_key="rk_test_...")
    try:
        tools = toolkit.get_tools()
        # ... use tools ...
    finally:
        await toolkit.close()
```

--------------------------------

### Initialize Stripe.js for Destination Charges

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/connect-recommend/references/charge-patterns.md

Standard frontend initialization using the platform's publishable key.

```javascript
// Initialize Stripe with platform's publishable key (no stripeAccount needed)
const stripe = await loadStripe('pk_test_platform_key');

const result = await stripe.confirmPayment({
  elements,
  confirmParams: {
    return_url: 'https://yoursite.com/success',
  },
});
```

--------------------------------

### Generate Solution and Product Catalog - Bash

Source: https://github.com/stripe/ai/blob/main/benchmarks/checkout-gym/README.md

Execute the provided script to create products in your Stripe account, generate the product catalog with new price IDs, and create the solution submission file.

```bash
./generate_solution.sh
```

--------------------------------

### Copy Environment Template

Source: https://github.com/stripe/ai/blob/main/tools/python/examples/crewai/README.md

Copy the .env.template file to .env and populate it with your specific values before running the application.

```bash
cp .env.template .env
```

--------------------------------

### Create Subscription for Partial Payments

Source: https://github.com/stripe/ai/blob/main/benchmarks/saas-starter-partial-payments/solution/SOLUTION_NOTES.md

Initiate a subscription with `payment_behavior: 'default_incomplete'` to enable manual payment handling. This sets up the subscription to await payment attachment.

```typescript
const subscription = await stripe.subscriptions.create({
  customer: customerId,
  items: [{ price: priceId }],
  payment_behavior: 'default_incomplete',
  expand: ['latest_invoice'],
});

const invoice = subscription.latest_invoice;
const totalAmount = invoice.amount_due;
const halfAmount = Math.floor(totalAmount / 2);
```

--------------------------------

### Create .env for Docker

Source: https://github.com/stripe/ai/blob/main/benchmarks/subscription-gym/README.md

Create a .env file in the subscription-gym root directory containing your Stripe secret key for Dockerized execution.

```bash
cat > .env << 'EOF'
STRIPE_SECRET_KEY=sk_test_...
EOF
```

--------------------------------

### Attach Payment Intent to Invoice

Source: https://github.com/stripe/ai/blob/main/benchmarks/saas-starter-partial-payments/solution/SOLUTION_NOTES.md

After a checkout session completes, retrieve the session to get the `payment_intent` ID. Then, use `stripe.invoices.attachPayment` to link this payment to the subscription's invoice.

```typescript
// After checkout session completes (via webhook or polling)
const session = await stripe.checkout.sessions.retrieve(sessionId);
const paymentIntentId = session.payment_intent;

// Attach payment to the subscription's invoice
await stripe.invoices.attachPayment(invoiceId, {
  payment: {
    payment_intent: paymentIntentId,
  },
});
```

--------------------------------

### Next Actions Inquiry

Source: https://github.com/stripe/ai/blob/main/providers/codex/plugin/skills/connect-recommend/SKILL.md

Prompt to determine the user's next steps after implementation guidance.

```text
What would you like to do next?
```

--------------------------------

### Run Database Migrations and Seed

Source: https://github.com/stripe/ai/blob/main/benchmarks/saas-starter-embedded-checkout/environment/README.md

Apply database migrations and seed the database with a default user and team. The default user is test@test.com with password admin123.

```bash
pnpm db:migrate
pnpm db:seed
```

--------------------------------

### Inspect Project State

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/stripe-projects/SKILL.md

Use these commands to view the status, environment variables, and service catalog of a project.

```bash
stripe projects status --json
```

```bash
stripe projects env --json
```

```bash
stripe projects catalog --json
```

--------------------------------

### Create LangChain Agent with Stripe Tools

Source: https://github.com/stripe/ai/blob/main/tools/typescript/README.md

Retrieve tools from the initialized StripeAgentToolkit and use them to create a LangChain agent. This requires llm, tools, and prompt to be defined elsewhere.

```typescript
import {AgentExecutor, createStructuredChatAgent} from 'langchain/agents';

const tools = stripeAgentToolkit.getTools();

const agent = await createStructuredChatAgent({
  llm,
  tools,
  prompt,
});

const agentExecutor = new AgentExecutor({
  agent,
  tools,
});
```

--------------------------------

### Create Test Payment

Source: https://github.com/stripe/ai/blob/main/benchmarks/sdk-upgrades-dotnet/environment/charges-on-payment-intent/server/README.md

Initiate a test payment using a POST request to the `/create-test-payment` endpoint. This will return a `paymentIntentId` needed for subsequent tests.

```bash
curl -X POST http://localhost:4242/create-test-payment
```

--------------------------------

### Populate test account with fixtures

Source: https://github.com/stripe/ai/blob/main/providers/codex/plugin/README.md

Downloads the test data JSON and applies it to the test account using the Stripe CLI.

```bash
curl -sL https://raw.githubusercontent.com/stripe/ai/main/providers/codex/plugin/fixtures/test-data.json \
  -o /tmp/stripe-mcp-test-data.json \
&& npx @stripe/cli@latest --api-key ... fixtures /tmp/stripe-mcp-test-data.json
```

--------------------------------

### Initialize Stripe Agent Toolkit for LangChain

Source: https://github.com/stripe/ai/blob/main/tools/typescript/README.md

Instantiate the StripeAgentToolkit for use with LangChain. Ensure your STRIPE_SECRET_KEY is set in the environment.

```typescript
import {StripeAgentToolkit} from '@stripe/agent-toolkit/langchain';

const stripeAgentToolkit = new StripeAgentToolkit({
  secretKey: process.env.STRIPE_SECRET_KEY!,
  configuration: {},
});
```

--------------------------------

### Build and Debug Stripe MCP Server Locally

Source: https://github.com/stripe/ai/blob/main/tools/modelcontextprotocol/README.md

Build the Stripe MCP server using npm and then run it with the MCP Inspector for debugging. Replace YOUR_STRIPE_SECRET_KEY with your actual Stripe secret key.

```bash
npm run build

npx @modelcontextprotocol/inspector node dist/index.js --api-key=YOUR_STRIPE_SECRET_KEY
```

--------------------------------

### Environment Configuration for Stripe and OpenAI

Source: https://github.com/stripe/ai/blob/main/tools/python/examples/openai/web_search/README.md

Populate this .env file with your OpenAI API key and Stripe secret key, customer ID, and meter ID to enable the agent's functionality.

```bash
OPENAI_API_KEY=your_openai_api_key
STRIPE_SECRET_KEY=your_stripe_secret_key
STRIPE_CUSTOMER_ID=your_stripe_customer_id
STRIPE_METER=your_stripe_meter
```

--------------------------------

### Configure Grader Environment

Source: https://github.com/stripe/ai/blob/main/benchmarks/subscription-gym/README.md

Copy the server's .env file to the grader directory so the grader can access Stripe API keys.

```bash
cp environment/server/.env grader/.env
```

--------------------------------

### Create Stripe Provider for Multi-turn Conversations

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/provider/README.md

Initialize the Stripe provider and use it for multi-turn conversations. Ensure you have your Stripe API key and customer ID configured.

```typescript
import { createStripe } from '@stripe/ai-sdk/provider';
import { generateText } from 'ai';

const stripeLLM = createStripe({
  apiKey: process.env.STRIPE_API_KEY,
  customerId: 'cus_xxxxx',
});

const result = await generateText({
  model: stripe('openai/gpt-4.1'),
  messages: [
    { role: 'user', content: 'What is the capital of France?' },
    { role: 'assistant', content: 'The capital of France is Paris.' },
    { role: 'user', content: 'What is its population?' },
  ],
});

console.log(result.text);
```

--------------------------------

### Using different providers with meteredModel

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/meter/README.md

Demonstrates how to wrap models from different AI SDK providers (OpenAI, Anthropic, Google) with `meteredModel` for usage tracking.

```typescript
import { meteredModel } from '@stripe/ai-sdk/meter';
import { openai } from '@ai-sdk/openai';
import { anthropic } from '@ai-sdk/anthropic';
import { google } from '@ai-sdk/google';

const STRIPE_API_KEY = process.env.STRIPE_API_KEY;
const CUSTOMER_ID = 'cus_xxxxx';

// OpenAI
const gptModel = meteredModel(
  openai('gpt-4o-mini'),
  STRIPE_API_KEY,
  CUSTOMER_ID
);

// Anthropic
const claudeModel = meteredModel(
  anthropic('claude-3-5-haiku-20241022'),
  STRIPE_API_KEY,
  CUSTOMER_ID
);

// Google
const geminiModel = meteredModel(
  google('gemini-2.5-flash'),
  STRIPE_API_KEY,
  CUSTOMER_ID
);
```

--------------------------------

### Discovery Checklist Update

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/connect-recommend/SKILL.md

Use this checklist format to track progress through the discovery and recommendation phases.

```text
  [x] Learn about your business
  [x] Scan your project
  [ ] Recommend configuration + charge pattern
  [ ] Produce recommendation plan
```

--------------------------------

### Manage Project Variables

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/stripe-projects/SKILL.md

Commands for creating, updating, and deleting custom project variables.

```bash
stripe projects variables set <name> --env-key <ENV_KEY> --value <value>
```

```bash
stripe projects variables list --json
```

```bash
stripe projects variables delete <name> --yes
```

--------------------------------

### Create Stripe Provider Instance

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/provider/README.md

Create an instance of the Stripe AI SDK provider. Requires your Stripe API key and optionally accepts a default customer ID.

```typescript
import { createStripe } from '@stripe/ai-sdk/provider';

const stripeLLM = createStripe({
  apiKey: process.env.STRIPE_API_KEY,
  customerId: 'cus_xxxxx', // Optional default customer ID
});
```

--------------------------------

### Verify capability status for onboarding gating

Source: https://github.com/stripe/ai/blob/main/providers/grok/plugin/skills/connect-recommend/references/recommendation-template.md

Use these checks to ensure connected accounts have the necessary capabilities active before enabling payouts or transfers.

```javascript
stripe.v2.core.accounts.retrieve(id)
```

```javascript
configuration.merchant.capabilities.card_payments.status === 'active'
```

```javascript
configuration.recipient.capabilities.stripe_balance.stripe_transfers.status === 'active'
```

--------------------------------

### Initialize Stripe.js for Direct Charges

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/connect-recommend/references/charge-patterns.md

Initialize the Stripe instance with the connected account ID to process payments directly on that account.

```javascript
// Must initialize Stripe with connected account
const stripe = await loadStripe('pk_test_...', {
  stripeAccount: 'acct_connected_account_id',
});

// Then confirm payment as usual
const result = await stripe.confirmPayment({
  elements,
  confirmParams: {
    return_url: 'https://yoursite.com/success',
  },
});
```

--------------------------------

### Platform Fee Structure Questions

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/connect-recommend/references/discovery-questions.md

Prompts for gathering platform fee structure preferences from the user.

```text
How do you want to charge your platform fee?
```

```text
What's your platform fee?
```

--------------------------------

### Search documentation

Source: https://github.com/stripe/ai/blob/main/providers/claude/plugin/skills/stripe-docs/SKILL.md

Performs a keyword search across the Stripe documentation.

```bash
stripe docs search "payment intents"
```

--------------------------------

### Run Docker Container with Stripe Secret Key

Source: https://github.com/stripe/ai/blob/main/benchmarks/saas-starter-embedded-checkout/README.md

Build the Docker image and run the container, passing the Stripe secret key using the -e flag. The key is then injected into the application's environment.

```bash
docker build -t saas-starter-embedded-checkout .
docker run -e STRIPE_SECRET_KEY=sk_test_... saas-starter-embedded-checkout
```

--------------------------------

### Display Progress Checklist

Source: https://github.com/stripe/ai/blob/main/providers/codex/plugin/skills/connect-recommend/SKILL.md

Use this text to inform the user about the upcoming steps in the recommendation process.

```text
Here's what we'll do:

  [ ] Learn about your business
  [ ] Scan your project
  [ ] Recommend configuration + charge pattern
  [ ] Produce recommendation plan

Let's get started.
```

--------------------------------

### Run Docker Container with Stripe Secret Key

Source: https://github.com/stripe/ai/blob/main/benchmarks/saas-starter-partial-payments/README.md

Build the Docker image and run the container, passing the Stripe secret key as an environment variable. The key is then written to the app's .env file inside the container.

```bash
docker build -t saas-starter-partial-payments .
docker run -e STRIPE_SECRET_KEY=sk_test_... saas-starter-partial-payments
```

--------------------------------

### Basic Usage Pattern with meteredModel

Source: https://github.com/stripe/ai/blob/main/llm/ai-sdk/meter/examples/README.md

Demonstrates the basic pattern of wrapping an AI SDK model with `meteredModel` to enable automatic Stripe meter event reporting.

```typescript
import { meteredModel } from '@stripe/ai-sdk-billing-wrapper';
import { openai } from '@ai-sdk/openai';
import { generateText } from 'ai';

const model = meteredModel(
  openai('gpt-4o-mini'),
  process.env.STRIPE_API_KEY,
  'cus_xxxxx'
);

const { text } = await generateText({
  model,
  prompt: 'Hello!',
});
```

--------------------------------

### Run Grading Tests

Source: https://github.com/stripe/ai/blob/main/benchmarks/galtee-invoicing/README.md

Execute the grading script to test the submission.

```bash
cd grader
./grade.sh
```
Tokens:
Raw