<h2>The Problem</h2>
<p>Financial apps give you data, but they make you work for it. Want to know how much you spent on food last month? Open the app, navigate to transactions, set filters, scroll through results. Want to compare it to the previous month? Start over.</p>
<p>I thought: what if you could just ask? "How much did I spend on restaurants in November?" and get an answer instantly, right in Discord where I already spend time.</p>
<h2>What I Built</h2>
<p>A Discord bot that combines <a href="https://platform.openai.com/">OpenAI's GPT-4</a> with <a href="https://plaid.com/">Plaid's</a> financial APIs. You chat naturally, and the bot figures out which Plaid endpoints to call based on what you're asking. It handles account balances, transaction history, spending insights, and more.</p>
<h2>How It Works</h2>
<p>The interesting part is how GPT-4 decides which API to call. The bot doesn't use rigid keyword matching. Instead, it sends your question to GPT-4 along with descriptions of available Plaid functions. GPT-4 picks the right one and extracts parameters from your natural language query.</p>
<pre><code>sequenceDiagram
    participant U as User
    participant D as Discord Bot
    participant G as GPT-4
    participant P as Plaid API

    U->>D: "What were my biggest purchases last week?"
    D->>G: Message + Available Functions
    G->>G: Interpret intent
    G-->>D: Call get_transactions(start, end, sort=amount)
    D->>P: Fetch transactions
    P-->>D: Transaction data
    D->>G: Format response
    G-->>D: Natural language summary
    D-->>U: "Your top 3 purchases were..."
</code></pre>
<h3>Tech Stack</h3>
<table>
<thead>
<tr>
<th>Layer</th>
<th>Technology</th>
<th>Why I Chose It</th>
</tr>
</thead>
<tbody>
<tr>
<td>Bot Framework</td>
<td><a href="https://discord.js.org/">discord.js</a></td>
<td>Most mature Discord library</td>
</tr>
<tr>
<td>AI</td>
<td><a href="https://platform.openai.com/">OpenAI GPT-4</a></td>
<td>Function calling capability</td>
</tr>
<tr>
<td>Financial Data</td>
<td><a href="https://plaid.com/">Plaid</a></td>
<td>Comprehensive financial APIs</td>
</tr>
<tr>
<td>Language</td>
<td><a href="https://www.typescriptlang.org/">TypeScript</a></td>
<td>Type safety for API responses</td>
</tr>
</tbody>
</table>
<h2>What Building This Taught Me</h2>
<h3>1. Function calling changes everything</h3>
<p>Before <a href="https://platform.openai.com/docs/guides/function-calling">function calling</a>, I was trying to parse user intent with regex and keyword matching. It was brittle and frustrating.</p>
<p>OpenAI's function calling feature made the bot actually understand what users wanted. You describe your available functions, and GPT-4 decides which one to call with the right parameters.</p>
<pre><code>// Define available functions for GPT-4
const functions = [
  {
    name: 'get_transactions',
    description: "Get user's transactions for a date range",
    parameters: {
      type: 'object',
      properties: {
        start_date: { type: 'string', description: 'Start date (YYYY-MM-DD)' },
        end_date: { type: 'string', description: 'End date (YYYY-MM-DD)' },
        category: { type: 'string', description: 'Optional category filter' },
      },
    },
  },
]

// GPT-4 automatically extracts: "last week" → actual dates
// "biggest purchases" → sort by amount descending
</code></pre>
<p>Resources:</p>
<ul>
<li><a href="https://platform.openai.com/docs/guides/function-calling">OpenAI Function Calling Guide</a></li>
<li><a href="https://www.deeplearning.ai/short-courses/functions-tools-agents-langchain/">Building AI Agents</a></li>
</ul>
<h3>2. Security is non negotiable with financial data</h3>
<p>Financial data requires serious security. I implemented strict permission systems so only authorized users can access their own data. No shared access, no admin backdoors.</p>
<pre><code>// Every request validates ownership
const userId = interaction.user.id
const accessToken = await getPlaidToken(userId)

if (!accessToken) {
  return 'Please link your bank account first'
}

// User can only see their own data
const transactions = await plaid.getTransactions(accessToken)
</code></pre>
<p>Resources:</p>
<ul>
<li><a href="https://plaid.com/docs/link/best-practices/">Plaid Security Best Practices</a></li>
<li><a href="https://discordjs.guide/popular-topics/permissions.html">Discord.js Permissions Guide</a></li>
</ul>
<h3>3. Discord.js has hidden complexity</h3>
<p>The library is powerful but the documentation assumes you already know Discord's permission model. I spent hours debugging why commands weren't registering, only to find it was a permission scope issue.</p>
<p>The slash command registration process is particularly tricky. Guild commands update instantly but global commands take up to an hour.</p>
<p>Resources:</p>
<ul>
<li><a href="https://discordjs.guide/">discord.js Guide</a></li>
<li><a href="https://discord.com/developers/docs">Discord Developer Portal</a></li>
</ul>
<h3>4. Rate limits compound when you chain APIs</h3>
<p>Both OpenAI and Plaid have rate limits. When you chain them together, you need to handle both. I added request queuing and graceful degradation when limits are hit.</p>
<pre><code>// Simple rate limiter
const rateLimiter = {
  openai: new Bottleneck({ maxConcurrent: 3, minTime: 200 }),
  plaid: new Bottleneck({ maxConcurrent: 5, minTime: 100 })
};

// Wrap API calls
const response = await rateLimiter.openai.schedule(() =>
  openai.chat.completions.create(...)
);
</code></pre>
<p>Resources:</p>
<ul>
<li><a href="https://github.com/SGrondin/bottleneck">Bottleneck Rate Limiter</a></li>
<li><a href="https://plaid.com/docs/api/rate-limits/">Plaid Rate Limits</a></li>
</ul>
<h2>The Bigger Realization</h2>
<p>Building this bot taught me that AI is best when it's invisible. Users don't want to learn prompts or understand how the AI works. They just want to ask questions and get answers.</p>
<p>The magic isn't the AI. It's that you can talk to your bank account like a person.</p>
<h2>What I'd Do Differently</h2>
<p>Cache more aggressively. Every question triggers API calls, which adds latency and cost. Many financial queries could use cached data that's a few minutes old without losing usefulness.</p>
<h2>References</h2>
<ul>
<li><a href="https://discord.js.org/">discord.js Documentation</a></li>
<li><a href="https://platform.openai.com/docs/api-reference">OpenAI API Reference</a></li>
<li><a href="https://plaid.com/docs/">Plaid API Documentation</a></li>
<li><a href="https://www.typescriptlang.org/docs/">TypeScript Handbook</a></li>
</ul>
<h2>Links</h2>
<ul>
<li><strong>GitHub:</strong> <a href="https://github.com/Summary-Box/discord-bot-cp">discord-bot-cp</a></li>
</ul>


Projects.

Discord Bot: Bridging AI and Financial Data