Pine Labs Agent Toolkit

Build AI-powered payment agents with LangChain, Vercel AI SDK, and OpenAI Agents SDK.

The Pine Labs Agent Toolkit enables popular agent frameworks — LangChain, Vercel's AI SDK, and OpenAI's Agents SDK — to integrate with Pine Labs payment APIs through function calling. Build AI-powered payment agents that can create orders, process payments, and manage transactions programmatically.

Pre-requisites

• Create a Pine Labs merchant account and obtain your Client ID and Client Secret from the Pine Labs Developer Dashboard.
• Node.js v18 or later.

Installation

# npm
npm install @plural_pinelabs/agent-toolkit

# yarn
yarn add @plural_pinelabs/agent-toolkit

# pnpm
pnpm add @plural_pinelabs/agent-toolkit

Framework Support

| Framework | Import Path | Description | | ------------- | ----------------------------------------- | --------------------------------------- | | OpenAI | @plural_pinelabs/agent-toolkit/openai | For Chat Completions API and Agents SDK | | LangChain | @plural_pinelabs/agent-toolkit/langchain | For LangChain agents | | Vercel AI SDK | @plural_pinelabs/agent-toolkit/ai-sdk | For Vercel's AI SDK |

Available Tools

| Tool | Description | Status | | --------------- | --------------------------------------------------------- | ------------ | | createOrder | Create a new payment order | ✅ Available | | getOrder | Retrieve details of an existing order by order ID | ✅ Available | | cancelOrder | Cancel or terminate an order | ✅ Available | | createRefund | Initiate a refund for an order | ✅ Available |

OpenAI

The OpenAI integration supports both the Chat Completions API and the Agents SDK.

With OpenAI Agents SDK

import { Agent, run } from '@openai/agents';
import {
  PinelabsAgentToolkit,
  pinelabsEnvironment,
} from '@plural_pinelabs/agent-toolkit/openai';

const pinelabs = new PinelabsAgentToolkit(
  pinelabsEnvironment.UAT,               // or pinelabsEnvironment.PRODUCTION
  process.env.PINE_CLIENT_ID!,
  process.env.PINE_CLIENT_SECRET!
);

const agent = new Agent({
  name: 'Payment Agent',
  instructions: 'You are a helpful payment assistant.',
  model: 'gpt-4o',
  tools: pinelabs.getAgentTools(),
});

const result = await run(agent, 'Create an order for Rs. 500 for customer [email protected]');
console.log(result.finalOutput);

With Chat Completions API

import OpenAI from 'openai';
import {
  PinelabsAgentToolkit,
  pinelabsEnvironment,
} from '@plural_pinelabs/agent-toolkit/openai';

const openai = new OpenAI();

const pinelabs = new PinelabsAgentToolkit(
  pinelabsEnvironment.UAT,
  process.env.PINE_CLIENT_ID!,
  process.env.PINE_CLIENT_SECRET!
);

const response = await openai.chat.completions.create({
  model: 'gpt-4o',
  messages: [{ role: 'user', content: 'Create an order for Rs. 500' }],
  tools: pinelabs.getTools(),
});

for (const toolCall of response.choices[0].message.tool_calls ?? []) {
  const result = await pinelabs.handleToolCall(
    toolCall.function.name,
    JSON.parse(toolCall.function.arguments)
  );
  console.log(result);
}

LangChain

import { ChatOpenAI } from '@langchain/openai';
import { AgentExecutor, createOpenAIFunctionsAgent } from 'langchain/agents';
import { ChatPromptTemplate } from '@langchain/core/prompts';
import {
  PinelabsAgentToolkit,
  pinelabsEnvironment,
} from '@plural_pinelabs/agent-toolkit/langchain';

const pinelabs = new PinelabsAgentToolkit(
  pinelabsEnvironment.UAT,
  process.env.PINE_CLIENT_ID!,
  process.env.PINE_CLIENT_SECRET!
);

const tools = pinelabs.getTools(); // DynamicStructuredTool[]

const llm = new ChatOpenAI({ model: 'gpt-4o' });
const prompt = ChatPromptTemplate.fromMessages([
  ['system', 'You are a helpful payment assistant.'],
  ['human', '{input}'],
  ['placeholder', '{agent_scratchpad}'],
]);

const agent = await createOpenAIFunctionsAgent({ llm, tools, prompt });
const executor = new AgentExecutor({ agent, tools });

const result = await executor.invoke({ input: 'Create an order for Rs. 500' });
console.log(result.output);

Vercel AI SDK

import { generateText } from 'ai';
import { openai } from '@ai-sdk/openai';
import {
  PinelabsAgentToolkit,
  pinelabsEnvironment,
} from '@plural_pinelabs/agent-toolkit/ai-sdk';

const pinelabs = new PinelabsAgentToolkit(
  pinelabsEnvironment.UAT,
  process.env.PINE_CLIENT_ID!,
  process.env.PINE_CLIENT_SECRET!
);

const result = await generateText({
  model: openai('gpt-4o'),
  tools: pinelabs.getTools(),
  prompt: 'Create an order for Rs. 500',
  maxSteps: 5,
});

console.log(result.text);

Environment Variables

| Variable | Description | | -------------------- | ------------------------------------------------ | | PINE_CLIENT_ID | Client ID from the Pine Labs Developer Dashboard | | PINE_CLIENT_SECRET | Client Secret from the Pine Labs Developer Dashboard |

Environments

| Environment | Constant | Base URL | | ----------- | -------------------------------- | ---------------------------------- | | UAT | pinelabsEnvironment.UAT | https://pluraluat.v2.pinepg.in | | Production | pinelabsEnvironment.PRODUCTION | https://api.pluralpay.in |

Authentication

All tools authenticate automatically using the client_id and client_secret passed to the toolkit constructor. Tokens are cached and refreshed before expiry with no manual management required.

For createOrder, the agent supplies the merchant client_id and client_secret directly as tool parameters alongside the order details. The SDK obtains a Bearer token internally and calls the checkout API in a single step.

Resources

• Pine Labs Developer Documentation
• npm Package
• GitHub Repository

License

MIT