Stripe Agent Toolkit - TypeScript

The Stripe Agent Toolkit enables popular agent frameworks including LangChain and Vercel's AI SDK to integrate with Stripe APIs through function calling. It also provides tooling to quickly integrate metered billing for prompt and completion token usage.

Installation

You don't need this source code unless you want to modify the package. If you just
want to use the package run:

``
npm install @stripe/agent-toolkit
`

$3

- Node 18+

Usage

The library needs to be configured with your account's secret key which is available in your [Stripe Dashboard][api-keys]. Additionally, configuration enables you to specify the types of actions that can be taken using the toolkit.

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

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

$3

The toolkit works with LangChain and Vercel's AI SDK and can be passed as a list of tools. For example:

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

const tools = stripeAgentToolkit.getTools();

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

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

#### Context

In some cases you will want to provide values that serve as defaults when making requests. Currently, the account context value enables you to make API calls for your connected accounts.

`typescript
const stripeAgentToolkit = new StripeAgentToolkit({
secretKey: process.env.STRIPE_SECRET_KEY!,
configuration: {
context: {
account: 'acct_123',
},
},
});
`

$3

For Vercel's AI SDK, you can use middleware to submit billing events for usage. All that is required is the customer ID and the input/output meters to bill.

`typescript
import {StripeAgentToolkit} from '@stripe/agent-toolkit/ai-sdk';
import {openai} from '@ai-sdk/openai';
import {
generateText,
experimental_wrapLanguageModel as wrapLanguageModel,
} from 'ai';

const stripeAgentToolkit = new StripeAgentToolkit({
secretKey: process.env.STRIPE_SECRET_KEY!,
configuration: {
actions: {
paymentLinks: {
create: true,
},
},
},
});

const model = wrapLanguageModel({
model: openai('gpt-4o'),
middleware: stripeAgentToolkit.middleware({
billing: {
customer: 'cus_123',
meters: {
input: 'input_tokens',
output: 'output_tokens',
},
},
}),
});
`

This works with both generateText and generateStream from the Vercel AI SDK.

Model Context Protocol

The Stripe Agent Toolkit also supports the Model Context Protocol (MCP). See /examples/modelcontextprotocol for an example. The same configuration options are available, and the server can be run with all supported transports.

`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: {
actions: {
paymentLinks: {
create: true,
},
products: {
create: true,
},
prices: {
create: true,
},
},
},
});

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);
});
``

[node-sdk]: https://github.com/stripe/stripe-node
[api-keys]: https://dashboard.stripe.com/account/apikeys