# How to integrate Ynab MCP with LlamaIndex

```json
{
  "title": "How to integrate Ynab MCP with LlamaIndex",
  "toolkit": "Ynab",
  "toolkit_slug": "ynab",
  "framework": "LlamaIndex",
  "framework_slug": "llama-index",
  "url": "https://composio.dev/toolkits/ynab/framework/llama-index",
  "markdown_url": "https://composio.dev/toolkits/ynab/framework/llama-index.md",
  "updated_at": "2026-05-12T10:30:56.806Z"
}
```

## Introduction

This guide walks you through connecting Ynab to LlamaIndex using the Composio tool router. By the end, you'll have a working Ynab agent that can show your spending by category this week, add a new expense to groceries, list upcoming scheduled transactions through natural language commands.
This guide will help you understand how to give your LlamaIndex agent real control over a Ynab account through Composio's Ynab MCP server.
Before we dive in, let's take a quick look at the key ideas and tools involved.

## Also integrate Ynab with

- [OpenAI Agents SDK](https://composio.dev/toolkits/ynab/framework/open-ai-agents-sdk)
- [Claude Agent SDK](https://composio.dev/toolkits/ynab/framework/claude-agents-sdk)
- [Claude Code](https://composio.dev/toolkits/ynab/framework/claude-code)
- [Claude Cowork](https://composio.dev/toolkits/ynab/framework/claude-cowork)
- [Codex](https://composio.dev/toolkits/ynab/framework/codex)
- [OpenClaw](https://composio.dev/toolkits/ynab/framework/openclaw)
- [Hermes](https://composio.dev/toolkits/ynab/framework/hermes-agent)
- [CLI](https://composio.dev/toolkits/ynab/framework/cli)
- [Google ADK](https://composio.dev/toolkits/ynab/framework/google-adk)
- [LangChain](https://composio.dev/toolkits/ynab/framework/langchain)
- [Vercel AI SDK](https://composio.dev/toolkits/ynab/framework/ai-sdk)
- [Mastra AI](https://composio.dev/toolkits/ynab/framework/mastra-ai)
- [CrewAI](https://composio.dev/toolkits/ynab/framework/crew-ai)

## TL;DR

Here's what you'll learn:
- Set your OpenAI and Composio API keys
- Install LlamaIndex and Composio packages
- Create a Composio Tool Router session for Ynab
- Connect LlamaIndex to the Ynab MCP server
- Build a Ynab-powered agent using LlamaIndex
- Interact with Ynab through natural language

## What is LlamaIndex?

LlamaIndex is a data framework for building LLM applications. It provides tools for connecting LLMs to external data sources and services through agents and tools.
Key features include:
- ReAct Agent: Reasoning and acting pattern for tool-using agents
- MCP Tools: Native support for Model Context Protocol
- Context Management: Maintain conversation context across interactions
- Async Support: Built for async/await patterns

## What is the Ynab MCP server, and what's possible with it?

The Ynab MCP server is an implementation of the Model Context Protocol that connects your AI agent and assistants like Claude, Cursor, etc directly to your Ynab account. It provides structured and secure access to your budgeting data, so your agent can perform actions like tracking expenses, creating and adjusting budgets, fetching transaction history, and analyzing spending patterns on your behalf.
- Automated transaction tracking: Let your agent fetch, categorize, or summarize recent transactions to keep your spending up to date.
- Budget creation and adjustment: Instruct your agent to set up new budgets, update existing categories, or reallocate funds between envelopes as your needs change.
- Spending analysis and reporting: Have your agent analyze your expenses, highlight trends, or generate summaries to help you understand where your money goes.
- Goal monitoring and progress updates: Ask your agent to track progress toward savings or debt payoff goals, and notify you of milestones or needed adjustments.
- Account and balance synchronization: Enable your agent to pull up-to-date balances from multiple accounts, ensuring your budget always reflects the latest financial picture.

## Supported Tools

| Tool slug | Name | Description |
|---|---|---|
| `YNAB_CREATE_ACCOUNT` | Create a new account in a budget | Tool to create a new account in a budget. Use when you need to add a manual (non-linked) account to a specific budget. |
| `YNAB_CREATE_SCHEDULED_TRANSACTION` | Create Scheduled Transaction | Creates a scheduled (recurring) transaction in a YNAB budget. Use this to set up bills, recurring expenses, or income that repeats on a schedule. Requires an account_id (from YNAB_LIST_ACCOUNTS), a date, frequency, and amount in milliunits. |
| `YNAB_DELETE_SCHEDULED_TRANSACTION` | Delete Scheduled Transaction | Permanently deletes a scheduled transaction from a YNAB budget. Use this when a recurring transaction is no longer needed. This action cannot be undone. Obtain the scheduled_transaction_id from list_scheduled_transactions or get_scheduled_transaction_by_id before calling this. |
| `YNAB_GET_ACCOUNT_BY_ID` | Get Account by ID | Tool to retrieve a single account resource. Use when you need details of a specific account by its ID after determining the budget. |
| `YNAB_GET_BUDGET_BY_ID` | Get Budget By ID | Retrieve a complete budget export by ID, including all accounts, categories, payees, and transactions. Use this tool when you need comprehensive budget data. The response includes: - Budget metadata (name, date/currency formats) - All accounts with balances - Category groups and categories with budgeted amounts and balances - All payees - All transactions and scheduled transactions (optional, may be large) For incremental syncing, use last_knowledge_of_server to only fetch changes since your last request. Note: Currency amounts are in 'milliunits' format (divide by 1000 for standard units). |
| `YNAB_GET_BUDGET_MONTH` | Get Budget Month | Tool to retrieve a specific budget month. Use when you need detailed summary and category budgets for a given month after confirming the budget ID. Monetary amounts are returned in milliunits (divide by 1000 for standard currency). Response includes high-level aggregates and a nested category array within the month payload for per-category budget analysis. |
| `YNAB_GET_BUDGET_SETTINGS` | Get Budget Settings | Tool to retrieve budget-level settings. Use when you need to fetch currency and date formatting preferences for a specific budget. |
| `YNAB_GET_CATEGORY_BY_ID` | Get Category by ID | Retrieve a single YNAB budget category by its ID. Use this to get detailed category information including budgeted amounts, activity, balance, and goal settings. Amounts are returned for the current budget month (UTC). Requires both budget_id and category_id - use 'last-used' or 'default' for budget_id if you don't have a specific budget ID. |
| `YNAB_GET_MONTH_CATEGORY_BY_ID` | Get Month Category By ID | Tool to retrieve a single category for a specific budget month. Use after selecting the budget and month when you need month-specific category details. |
| `YNAB_GET_PAYEE_BY_ID` | Get Payee by ID | Tool to retrieve a single payee by its ID. Use when you need full details of a payee after confirming the budget and payee selection. |
| `YNAB_GET_SCHEDULED_TRANSACTION_BY_ID` | Get Scheduled Transaction by ID | Tool to retrieve a single scheduled transaction. Use when you need details of a specific scheduled transaction by its ID after confirming the budget. |
| `YNAB_GET_TRANSACTION_BY_ID` | Get Transaction by ID | Tool to retrieve a single transaction by its ID. Use when you need detailed information for a specific transaction after confirming the budget. |
| `YNAB_GET_USER` | Get Authorized User | Tool to retrieve the authorized user's information. Use after authentication to get current user details. |
| `YNAB_LIST_ACCOUNTS` | List Accounts | Retrieves all accounts for a specified YNAB budget. Use this tool to: - Get a complete list of accounts (checking, savings, credit cards, loans, etc.) in a budget - Retrieve account IDs for use with other YNAB account-related operations - Monitor account balances (current, cleared, and uncleared) - Perform efficient delta syncing by passing last_knowledge_of_server Returns account details including balances in milliunits (1000 milliunits = 1 currency unit). For example, a balance of 150000 milliunits equals $150.00. Note: Use 'default' or 'last-used' as budget_id for convenience, or obtain specific budget UUIDs from the List Budgets endpoint. |
| `YNAB_LIST_BUDGETS` | List Budgets | Retrieves all budgets accessible to the authenticated YNAB user. Use this tool to: - Get a list of all budgets in the user's YNAB account - Retrieve budget IDs for use with other YNAB operations - Optionally include account details for each budget Returns budget summaries including name, last modification time, date range, and formatting preferences. Budget IDs from this response are required for most other YNAB API operations. |
| `YNAB_LIST_CATEGORIES` | List Categories | Tool to list all category groups and their categories for a budget. Use when you need the full set of budget categories or only those changed since last sync. Category names may include emojis or near-identical labels; always use category_id from the response in downstream requests. |
| `YNAB_LIST_MONTHS` | List Budget Months | Tool to list months for a budget with summary information. Use when you need an overview of all months or only changed months since a given knowledge value (delta requests). |
| `YNAB_LIST_PAYEE_LOCATIONS` | List Payee Locations | Retrieves all GPS locations associated with payees in a budget. Payee locations are recorded when transactions are saved using the YNAB mobile apps, storing GPS coordinates to help pre-populate nearby payees. Use this tool when you need to find all recorded locations for payees in a budget. Note: Not all payees will have locations recorded. |
| `YNAB_LIST_PAYEE_LOCATIONS_BY_PAYEE` | List Payee Locations by Payee | Retrieves all GPS locations associated with a specific payee in a budget. Payee locations are recorded when transactions are saved using the YNAB mobile apps, storing GPS coordinates to help pre-populate nearby payees. Use this tool when you need to find where a payee's transactions typically occur. Note: Not all payees will have locations recorded. |
| `YNAB_LIST_PAYEES` | List payees in a budget | Tool to list payees in a budget. Use when you need all payees or only those changed since a known server knowledge value. |
| `YNAB_LIST_SCHEDULED_TRANSACTIONS` | List Scheduled Transactions | List all scheduled (recurring) transactions for a YNAB budget. Returns future-dated transactions that repeat on a schedule (e.g., rent, subscriptions, paychecks). Supports delta sync to fetch only changes since a previous request using the server_knowledge value. |
| `YNAB_LIST_TRANSACTIONS` | List Transactions | Tool to list transactions in a budget. Use when you need all or delta-filtered transactions after confirming the budget. Omitting all filters returns the full transaction history, which may produce large payloads; prefer since_date, type, or last_knowledge_of_server to scope results. |
| `YNAB_LIST_TRANSACTIONS_BY_ACCOUNT` | List Transactions by Account | Tool to list transactions for a specific account. Use when you need all or delta-filtered transactions for a given account after confirming the budget and account ID. |
| `YNAB_LIST_TRANSACTIONS_BY_CATEGORY` | List Transactions by Category | Tool to list transactions for a specific category (excludes pending). Use when you need completed transactions for a given category after obtaining budget_id and category_id. Response includes deleted transactions; filter client-side using the deleted boolean field. |
| `YNAB_LIST_TRANSACTIONS_BY_PAYEE` | List Transactions by Payee | Tool to list transactions for a specific payee in a budget. Use when you need to fetch all non-pending transactions for a payee after confirming budget and payee IDs. |
| `YNAB_UPDATE_PAYEE` | Update Payee | Tool to update a payee. Use when you need to rename an existing payee. |
| `YNAB_UPDATE_SCHEDULED_TRANSACTION` | Update Scheduled Transaction | Update an existing scheduled transaction in YNAB. Allows modifying the date, frequency, amount, payee, category, memo, or account of a scheduled transaction. Use YNAB_LIST_SCHEDULED_TRANSACTIONS first to get the scheduled_transaction_id. Amounts should be in milliunits (negative for expenses, positive for income). |

## Supported Triggers

None listed.

## Creating MCP Server - Stand-alone vs Composio SDK

The Ynab MCP server is an implementation of the Model Context Protocol that connects your AI agent to Ynab. It provides structured and secure access so your agent can perform Ynab operations on your behalf through a secure, permission-based interface.
With Composio's managed implementation, you don't have to create your own developer app. For production, if you're building an end product, we recommend using your own credentials. The managed server helps you prototype fast and go from 0-1 faster.

## Step-by-step Guide

### 1. Prerequisites

Before you begin, make sure you have:
- Python 3.8/Node 16 or higher installed
- A Composio account with the API key
- An OpenAI API key
- A Ynab account and project
- Basic familiarity with async Python/Typescript

### 1. Getting API Keys for OpenAI, Composio, and Ynab

No description provided.

### 2. Installing dependencies

No description provided.
```python
pip install composio-llamaindex llama-index llama-index-llms-openai llama-index-tools-mcp python-dotenv
```

```typescript
npm install @composio/llamaindex @llamaindex/openai @llamaindex/tools @llamaindex/workflow dotenv
```

### 3. Set environment variables

Create a .env file in your project root:
These credentials will be used to:
- Authenticate with OpenAI's GPT-5 model
- Connect to Composio's Tool Router
- Identify your Composio user session for Ynab access
```bash
OPENAI_API_KEY=your-openai-api-key
COMPOSIO_API_KEY=your-composio-api-key
COMPOSIO_USER_ID=your-user-id
```

### 4. Import modules

No description provided.
```python
import asyncio
import os
import dotenv

from composio import Composio
from composio_llamaindex import LlamaIndexProvider
from llama_index.core.agent.workflow import ReActAgent
from llama_index.core.workflow import Context
from llama_index.llms.openai import OpenAI
from llama_index.tools.mcp import BasicMCPClient, McpToolSpec

dotenv.load_dotenv()
```

```typescript
import "dotenv/config";
import readline from "node:readline/promises";
import { stdin as input, stdout as output } from "node:process";

import { Composio } from "@composio/core";

import { mcp } from "@llamaindex/tools";
import { agent as createAgent } from "@llamaindex/workflow";
import { openai } from "@llamaindex/openai";

dotenv.config();
```

### 5. Load environment variables and initialize Composio

No description provided.
```python
OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
COMPOSIO_API_KEY = os.getenv("COMPOSIO_API_KEY")
COMPOSIO_USER_ID = os.getenv("COMPOSIO_USER_ID")

if not OPENAI_API_KEY:
    raise ValueError("OPENAI_API_KEY is not set in the environment")
if not COMPOSIO_API_KEY:
    raise ValueError("COMPOSIO_API_KEY is not set in the environment")
if not COMPOSIO_USER_ID:
    raise ValueError("COMPOSIO_USER_ID is not set in the environment")
```

```typescript
const OPENAI_API_KEY = process.env.OPENAI_API_KEY;
const COMPOSIO_API_KEY = process.env.COMPOSIO_API_KEY;
const COMPOSIO_USER_ID = process.env.COMPOSIO_USER_ID;

if (!OPENAI_API_KEY) throw new Error("OPENAI_API_KEY is not set");
if (!COMPOSIO_API_KEY) throw new Error("COMPOSIO_API_KEY is not set");
if (!COMPOSIO_USER_ID) throw new Error("COMPOSIO_USER_ID is not set");
```

### 6. Create a Tool Router session and build the agent function

What's happening here:
- We create a Composio client using your API key and configure it with the LlamaIndex provider
- We then create a tool router MCP session for your user, specifying the toolkits we want to use (in this case, ynab)
- The session returns an MCP HTTP endpoint URL that acts as a gateway to all your configured tools
- LlamaIndex will connect to this endpoint to dynamically discover and use the available Ynab tools.
- The MCP tools are mapped to LlamaIndex-compatible tools and plug them into the Agent.
```python
async def build_agent() -> ReActAgent:
    composio_client = Composio(
        api_key=COMPOSIO_API_KEY,
        provider=LlamaIndexProvider(),
    )

    session = composio_client.create(
        user_id=COMPOSIO_USER_ID,
        toolkits=["ynab"],
    )

    mcp_url = session.mcp.url
    print(f"Composio MCP URL: {mcp_url}")

    mcp_client = BasicMCPClient(mcp_url, headers={"x-api-key": COMPOSIO_API_KEY})
    mcp_tool_spec = McpToolSpec(client=mcp_client)
    tools = await mcp_tool_spec.to_tool_list_async()

    llm = OpenAI(model="gpt-5")

    description = "An agent that uses Composio Tool Router MCP tools to perform Ynab actions."
    system_prompt = """
    You are a helpful assistant connected to Composio Tool Router.
    Use the available tools to answer user queries and perform Ynab actions.
    """
    return ReActAgent(tools=tools, llm=llm, description=description, system_prompt=system_prompt, verbose=True)
```

```typescript
async function buildAgent() {

  console.log(`Initializing Composio client...${COMPOSIO_USER_ID!}...`);
  console.log(`COMPOSIO_USER_ID: ${COMPOSIO_USER_ID!}...`);

  const composio = new Composio({
    apiKey: COMPOSIO_API_KEY,
    provider: new LlamaindexProvider(),
  });

  const session = await composio.create(
    COMPOSIO_USER_ID!,
    {
      toolkits: ["ynab"],
    },
  );

  const mcpUrl = session.mcp.url;
  console.log(`Composio Tool Router MCP URL: ${mcpUrl}`);

  const server = mcp({
    url: mcpUrl,
    clientName: "composio_tool_router_with_llamaindex",
    requestInit: {
      headers: {
        "x-api-key": COMPOSIO_API_KEY!,
      },
    },
    // verbose: true,
  });

  const tools = await server.tools();

  const llm = openai({ apiKey: OPENAI_API_KEY, model: "gpt-5" });

  const agent = createAgent({
    name: "composio_tool_router_with_llamaindex",
        description : "An agent that uses Composio Tool Router MCP tools to perform actions.",
    systemPrompt:
      "You are a helpful assistant connected to Composio Tool Router."+
"Use the available tools to answer user queries and perform Ynab actions." ,
    llm,
    tools,
  });

  return agent;
}
```

### 7. Create an interactive chat loop

No description provided.
```python
async def chat_loop(agent: ReActAgent) -> None:
    ctx = Context(agent)
    print("Type 'quit', 'exit', or Ctrl+C to stop.")

    while True:
        try:
            user_input = input("\nYou: ").strip()
        except (KeyboardInterrupt, EOFError):
            print("\nBye!")
            break

        if not user_input or user_input.lower() in {"quit", "exit"}:
            print("Bye!")
            break

        try:
            print("Agent: ", end="", flush=True)
            handler = agent.run(user_input, ctx=ctx)

            async for event in handler.stream_events():
                # Stream token-by-token from LLM responses
                if hasattr(event, "delta") and event.delta:
                    print(event.delta, end="", flush=True)
                # Show tool calls as they happen
                elif hasattr(event, "tool_name"):
                    print(f"\n[Using tool: {event.tool_name}]", flush=True)

            # Get final response
            response = await handler
            print()  # Newline after streaming
        except KeyboardInterrupt:
            print("\n[Interrupted]")
            continue
        except Exception as e:
            print(f"\nError: {e}")
```

```typescript
async function chatLoop(agent: ReturnType<typeof createAgent>) {
  const rl = readline.createInterface({ input, output });

  console.log("Type 'quit' or 'exit' to stop.");

  while (true) {
    let userInput: string;

    try {
      userInput = (await rl.question("\nYou: ")).trim();
    } catch {
      console.log("\nAgent: Bye!");
      break;
    }

    if (!userInput) {
      continue;
    }

    const lower = userInput.toLowerCase();
    if (lower === "quit" || lower === "exit") {
      console.log("Agent: Bye!");
      break;
    }

    try {
      process.stdout.write("Agent: ");

      const stream = agent.runStream(userInput);
      let finalResult: any = null;

      for await (const event of stream) {
        // The event.data contains the streamed content
        const data: any = event.data;

        // Check for streaming delta content
        if (data?.delta) {
          process.stdout.write(data.delta);
        }

        // Store final result for fallback
        if (data?.result || data?.message) {
          finalResult = data;
        }
      }

      // If no streaming happened, show the final result
      if (finalResult) {
        const answer =
          finalResult.result ??
          finalResult.message?.content ??
          finalResult.message ??
          "";
        if (answer && typeof answer === "string" && !answer.includes("[object")) {
          process.stdout.write(answer);
        }
      }

      console.log(); // New line after streaming completes
    } catch (err: any) {
      console.error("\nAgent error:", err?.message ?? err);
    }
  }

  rl.close();
}
```

### 8. Define the main entry point

What's happening here:
- We're orchestrating the entire application flow
- The agent gets built with proper error handling
- Then we kick off the interactive chat loop so you can start talking to Ynab
```python
async def main() -> None:
    agent = await build_agent()
    await chat_loop(agent)

if __name__ == "__main__":
    # Handle Ctrl+C gracefully
    signal.signal(signal.SIGINT, lambda s, f: (print("\nBye!"), exit(0)))
    try:
        asyncio.run(main())
    except KeyboardInterrupt:
        print("\nBye!")
```

```typescript
async function main() {
  try {
    const agent = await buildAgent();
    await chatLoop(agent);
  } catch (err) {
    console.error("Failed to start agent:", err);
    process.exit(1);
  }
}

main();
```

### 9. Run the agent

When prompted, authenticate and authorise your agent with Ynab, then start asking questions.
```bash
python llamaindex_agent.py
```

```typescript
npx ts-node llamaindex-agent.ts
```

## Complete Code

```python
import asyncio
import os
import signal
import dotenv

from composio import Composio
from composio_llamaindex import LlamaIndexProvider
from llama_index.core.agent.workflow import ReActAgent
from llama_index.core.workflow import Context
from llama_index.llms.openai import OpenAI
from llama_index.tools.mcp import BasicMCPClient, McpToolSpec

dotenv.load_dotenv()

OPENAI_API_KEY = os.getenv("OPENAI_API_KEY")
COMPOSIO_API_KEY = os.getenv("COMPOSIO_API_KEY")
COMPOSIO_USER_ID = os.getenv("COMPOSIO_USER_ID")

if not OPENAI_API_KEY:
    raise ValueError("OPENAI_API_KEY is not set")
if not COMPOSIO_API_KEY:
    raise ValueError("COMPOSIO_API_KEY is not set")
if not COMPOSIO_USER_ID:
    raise ValueError("COMPOSIO_USER_ID is not set")

async def build_agent() -> ReActAgent:
    composio_client = Composio(
        api_key=COMPOSIO_API_KEY,
        provider=LlamaIndexProvider(),
    )

    session = composio_client.create(
        user_id=COMPOSIO_USER_ID,
        toolkits=["ynab"],
    )

    mcp_url = session.mcp.url
    print(f"Composio MCP URL: {mcp_url}")

    mcp_client = BasicMCPClient(mcp_url, headers={"x-api-key": COMPOSIO_API_KEY})
    mcp_tool_spec = McpToolSpec(client=mcp_client)
    tools = await mcp_tool_spec.to_tool_list_async()

    llm = OpenAI(model="gpt-5")
    description = "An agent that uses Composio Tool Router MCP tools to perform Ynab actions."
    system_prompt = """
    You are a helpful assistant connected to Composio Tool Router.
    Use the available tools to answer user queries and perform Ynab actions.
    """
    return ReActAgent(
        tools=tools,
        llm=llm,
        description=description,
        system_prompt=system_prompt,
        verbose=True,
    );

async def chat_loop(agent: ReActAgent) -> None:
    ctx = Context(agent)
    print("Type 'quit', 'exit', or Ctrl+C to stop.")

    while True:
        try:
            user_input = input("\nYou: ").strip()
        except (KeyboardInterrupt, EOFError):
            print("\nBye!")
            break

        if not user_input or user_input.lower() in {"quit", "exit"}:
            print("Bye!")
            break

        try:
            print("Agent: ", end="", flush=True)
            handler = agent.run(user_input, ctx=ctx)

            async for event in handler.stream_events():
                # Stream token-by-token from LLM responses
                if hasattr(event, "delta") and event.delta:
                    print(event.delta, end="", flush=True)
                # Show tool calls as they happen
                elif hasattr(event, "tool_name"):
                    print(f"\n[Using tool: {event.tool_name}]", flush=True)

            # Get final response
            response = await handler
            print()  # Newline after streaming
        except KeyboardInterrupt:
            print("\n[Interrupted]")
            continue
        except Exception as e:
            print(f"\nError: {e}")

async def main() -> None:
    agent = await build_agent()
    await chat_loop(agent)

if __name__ == "__main__":
    # Handle Ctrl+C gracefully
    signal.signal(signal.SIGINT, lambda s, f: (print("\nBye!"), exit(0)))
    try:
        asyncio.run(main())
    except KeyboardInterrupt:
        print("\nBye!")
```

```typescript
import "dotenv/config";
import readline from "node:readline/promises";
import { stdin as input, stdout as output } from "node:process";

import { Composio } from "@composio/core";
import { LlamaindexProvider } from "@composio/llamaindex";

import { mcp } from "@llamaindex/tools";
import { agent as createAgent } from "@llamaindex/workflow";
import { openai } from "@llamaindex/openai";

dotenv.config();

const OPENAI_API_KEY = process.env.OPENAI_API_KEY;
const COMPOSIO_API_KEY = process.env.COMPOSIO_API_KEY;
const COMPOSIO_USER_ID = process.env.COMPOSIO_USER_ID;

if (!OPENAI_API_KEY) {
    throw new Error("OPENAI_API_KEY is not set in the environment");
  }
if (!COMPOSIO_API_KEY) {
    throw new Error("COMPOSIO_API_KEY is not set in the environment");
  }
if (!COMPOSIO_USER_ID) {
    throw new Error("COMPOSIO_USER_ID is not set in the environment");
  }

async function buildAgent() {

  console.log(`Initializing Composio client...${COMPOSIO_USER_ID!}...`);
  console.log(`COMPOSIO_USER_ID: ${COMPOSIO_USER_ID!}...`);

  const composio = new Composio({
    apiKey: COMPOSIO_API_KEY,
    provider: new LlamaindexProvider(),
  });

  const session = await composio.create(
    COMPOSIO_USER_ID!,
    {
      toolkits: ["ynab"],
    },
  );

  const mcpUrl = session.mcp.url;
  console.log(`Composio Tool Router MCP URL: ${mcpUrl}`);

  const server = mcp({
    url: mcpUrl,
    clientName: "composio_tool_router_with_llamaindex",
    requestInit: {
      headers: {
        "x-api-key": COMPOSIO_API_KEY!,
      },
    },
    // verbose: true,
  });

  const tools = await server.tools();

  const llm = openai({ apiKey: OPENAI_API_KEY, model: "gpt-5" });

  const agent = createAgent({
    name: "composio_tool_router_with_llamaindex",
    description:
      "An agent that uses Composio Tool Router MCP tools to perform actions.",
    systemPrompt:
      "You are a helpful assistant connected to Composio Tool Router."+
"Use the available tools to answer user queries and perform Ynab actions." ,
    llm,
    tools,
  });

  return agent;
}

async function chatLoop(agent: ReturnType<typeof createAgent>) {
  const rl = readline.createInterface({ input, output });

  console.log("Type 'quit' or 'exit' to stop.");

  while (true) {
    let userInput: string;

    try {
      userInput = (await rl.question("\nYou: ")).trim();
    } catch {
      console.log("\nAgent: Bye!");
      break;
    }

    if (!userInput) {
      continue;
    }

    const lower = userInput.toLowerCase();
    if (lower === "quit" || lower === "exit") {
      console.log("Agent: Bye!");
      break;
    }

    try {
      process.stdout.write("Agent: ");

      const stream = agent.runStream(userInput);
      let finalResult: any = null;

      for await (const event of stream) {
        // The event.data contains the streamed content
        const data: any = event.data;

        // Check for streaming delta content
        if (data?.delta) {
          process.stdout.write(data.delta);
        }

        // Store final result for fallback
        if (data?.result || data?.message) {
          finalResult = data;
        }
      }

      // If no streaming happened, show the final result
      if (finalResult) {
        const answer =
          finalResult.result ??
          finalResult.message?.content ??
          finalResult.message ??
          "";
        if (answer && typeof answer === "string" && !answer.includes("[object")) {
          process.stdout.write(answer);
        }
      }

      console.log(); // New line after streaming completes
    } catch (err: any) {
      console.error("\nAgent error:", err?.message ?? err);
    }
  }

  rl.close();
}

async function main() {
  try {
    const agent = await buildAgent();
    await chatLoop(agent);
  } catch (err: any) {
    console.error("Failed to start agent:", err?.message ?? err);
    process.exit(1);
  }
}

main();
```

## Conclusion

You've successfully connected Ynab to LlamaIndex through Composio's Tool Router MCP layer.
Key takeaways:
- Tool Router dynamically exposes Ynab tools through an MCP endpoint
- LlamaIndex's ReActAgent handles reasoning and orchestration; Composio handles integrations
- The agent becomes more capable without increasing prompt size
- Async Python provides clean, efficient execution of agent workflows
You can easily extend this to other toolkits like Gmail, Notion, Stripe, GitHub, and more by adding them to the toolkits parameter.

## How to build Ynab MCP Agent with another framework

- [OpenAI Agents SDK](https://composio.dev/toolkits/ynab/framework/open-ai-agents-sdk)
- [Claude Agent SDK](https://composio.dev/toolkits/ynab/framework/claude-agents-sdk)
- [Claude Code](https://composio.dev/toolkits/ynab/framework/claude-code)
- [Claude Cowork](https://composio.dev/toolkits/ynab/framework/claude-cowork)
- [Codex](https://composio.dev/toolkits/ynab/framework/codex)
- [OpenClaw](https://composio.dev/toolkits/ynab/framework/openclaw)
- [Hermes](https://composio.dev/toolkits/ynab/framework/hermes-agent)
- [CLI](https://composio.dev/toolkits/ynab/framework/cli)
- [Google ADK](https://composio.dev/toolkits/ynab/framework/google-adk)
- [LangChain](https://composio.dev/toolkits/ynab/framework/langchain)
- [Vercel AI SDK](https://composio.dev/toolkits/ynab/framework/ai-sdk)
- [Mastra AI](https://composio.dev/toolkits/ynab/framework/mastra-ai)
- [CrewAI](https://composio.dev/toolkits/ynab/framework/crew-ai)

## Related Toolkits

- [Stripe](https://composio.dev/toolkits/stripe) - Stripe is a global online payments platform offering APIs for managing payments, customers, and subscriptions. Trusted by businesses for secure, efficient, and scalable payment processing worldwide.
- [Alpha vantage](https://composio.dev/toolkits/alpha_vantage) - Alpha Vantage is a financial data platform offering real-time and historical stock market APIs. Get instant, reliable access to equities, forex, and technical analysis data for smarter trading decisions.
- [Altoviz](https://composio.dev/toolkits/altoviz) - Altoviz is a cloud-based billing and invoicing platform for businesses. It streamlines online payments, expense tracking, and customizable invoice management.
- [Benzinga](https://composio.dev/toolkits/benzinga) - Benzinga provides real-time financial news and data APIs for market coverage. It helps you track breaking news and actionable market insights instantly.
- [Brex](https://composio.dev/toolkits/brex) - Brex provides corporate credit cards and spend management tailored for startups and tech businesses. It helps optimize company cash flow, streamline accounting, and accelerate business growth.
- [Chaser](https://composio.dev/toolkits/chaser) - Chaser is accounts receivable automation software that sends invoice reminders and helps businesses get paid faster. It streamlines the collections process to save time and improve cash flow.
- [Clientary](https://composio.dev/toolkits/clientary) - Clientary is a platform for managing clients, invoices, projects, proposals, and more. It streamlines client work and saves you serious admin time.
- [Coinbase](https://composio.dev/toolkits/coinbase) - Coinbase is a platform for buying, selling, and storing cryptocurrency. It makes exchanging and managing crypto simple and secure for everyone.
- [Coinranking](https://composio.dev/toolkits/coinranking) - Coinranking is a comprehensive cryptocurrency market data platform offering access to real-time coin prices, market caps, and historical data. Get accurate, up-to-date stats for thousands of digital assets in one place.
- [Coupa](https://composio.dev/toolkits/coupa) - Coupa is a business spend management platform for procurement, invoicing, and expenses. It helps organizations streamline purchasing, control costs, and gain complete visibility over financial operations.
- [CurrencyScoop](https://composio.dev/toolkits/currencyscoop) - CurrencyScoop is a developer-friendly API for real-time and historical currency exchange rates. Easily access fiat and crypto data for smart, up-to-date financial applications.
- [Daffy](https://composio.dev/toolkits/daffy) - Daffy is a modern charitable giving platform with a donor-advised fund. Easily set aside funds, grow them tax-free, and donate to over 1.7 million U.S. charities.
- [Eagle doc](https://composio.dev/toolkits/eagle_doc) - Eagle doc is an AI-powered OCR API for invoices and receipts. It delivers fast, reliable, and accurate document data extraction for seamless automation.
- [Elorus](https://composio.dev/toolkits/elorus) - Elorus is an online invoicing and time-tracking software for freelancers and small businesses. Easily manage finances, bill clients, and track work in one place.
- [Eodhd apis](https://composio.dev/toolkits/eodhd_apis) - Eodhd apis delivers comprehensive financial data, including live and historical stock prices, via robust APIs. Easily access reliable, up-to-date market insights to power your apps, dashboards, and analytics.
- [Fidel api](https://composio.dev/toolkits/fidel_api) - Fidel api is a secure platform for linking payment cards to web and mobile apps. It enables real-time card transaction monitoring and event-based automation for businesses.
- [Finage](https://composio.dev/toolkits/finage) - Finage is a secure API platform delivering real-time and historical financial data for stocks, forex, crypto, indices, and commodities. It empowers developers and businesses to access, analyze, and act on market data instantly.
- [Finmei](https://composio.dev/toolkits/finmei) - Finmei is an invoicing tool that simplifies billing, invoice management, and expense tracking. Ideal for automating and organizing your business finances in one place.
- [Fixer](https://composio.dev/toolkits/fixer) - Fixer is a currency data API offering real-time and historical exchange rates for 170 currencies. Instantly access accurate, up-to-date forex data for your applications and workflows.
- [Fixer io](https://composio.dev/toolkits/fixer_io) - Fixer.io is a lightweight API for real-time and historical foreign exchange rates. It makes global currency conversion fast, accurate, and hassle-free.

## Frequently Asked Questions

### What are the differences in Tool Router MCP and Ynab MCP?

With a standalone Ynab MCP server, the agents and LLMs can only access a fixed set of Ynab tools tied to that server. However, with the Composio Tool Router, agents can dynamically load tools from Ynab and many other apps based on the task at hand, all through a single MCP endpoint.

### Can I use Tool Router MCP with LlamaIndex?

Yes, you can. LlamaIndex fully supports MCP integration. You get structured tool calling, message history handling, and model orchestration while Tool Router takes care of discovering and serving the right Ynab tools.

### Can I manage the permissions and scopes for Ynab while using Tool Router?

Yes, absolutely. You can configure which Ynab scopes and actions are allowed when connecting your account to Composio. You can also bring your own OAuth credentials or API configuration so you keep full control over what the agent can do.

### How safe is my data with Composio Tool Router?

All sensitive data such as tokens, keys, and configuration is fully encrypted at rest and in transit. Composio is SOC 2 Type 2 compliant and follows strict security practices so your Ynab data and credentials are handled as safely as possible.

---
[See all toolkits](https://composio.dev/toolkits) · [Composio docs](https://docs.composio.dev/llms.txt)