When Wealthsimple launched their new credit card, I was excited. Simple rewards, clean design, good app experience. But when I went to pull my credit card transactions to import into my personal finance tracker, I realized something was missing: There's no CSV export option.

At least, not yet. So I built my own.

This post covers how I built a quick browser script to export all my Wealthsimple Credit Card transactions, the weird problems I ran into, and how I integrated the data into my app.

Getting the Raw HTML

The first step was figuring out how the transaction list was actually structured.

• I logged into my Wealthsimple account.
• Opened the Credit Card section.
• Right-clicked on the transactions page and selected Inspect Element to open the Developer Tools. Yes, I use Safari on my personal laptop. If you don't see that option, you'll have to turn on Developer tools — or use Chrome, Firefox, etc 😄.
• Inspected the HTML to see what was going on.

I found that Wealthsimple uses random class names (like .sc-e8c84276-1), meaning I couldn’t rely on CSS selectors. Instead, I needed to target the HTML structure (e.g., <h2> for dates, <button> for transactions).

Scraping HTML like this is always prone to breakage, but it will work for now!

Writing the Extraction Script

I wrote a quick JavaScript script that:

• Walks through the DOM.
• Picks up the date, description, and amount for each transaction.
• Downloads everything as a .csv file.

At first the results were messy:

• Duplicate transactions.
• Refunds and purchases were not distinguished.
• Fields with commas (like April 23, 2025) broke the CSV parser.

After a few iterations, I ended up with a cleaner, more robust script. Quoting fields correctly in CSVs is crucial when your data contains commas. See the Appendix for the full script.

Hidden Problems: Unicode Minus Signs

One issue was sneaky: Wealthsimple’s UI uses a Unicode minus character (U+2212) instead of the normal ASCII -.

This meant when I parsed the amounts in Ruby using .to_f, negative numbers were silently treated as positive!

Example:

• "−73.97" (Unicode minus) → incorrectly parsed as 73.97

The fix was easy enough once I realized the issue — normalize the minus sign in Ruby before parsing:

def parse_amount(amount_str)
  cleaned = amount_str.tr("−", "-").gsub(/[^\d\.\-]/, '')
  amount = cleaned.to_f
  (-amount * 100).round
end

Integrating With My App

I already had an importer service that handles credit card transactions in .csv files from other banks. Adding Wealthsimple support was just a matter of creating a new importer class that:

• Cleans the date and amount fields properly.
• Inverts the transaction value as my app uses negative for refunds.
• Import everything as transactions into the database.

My app code isn’t public yet, but it’s built around a flexible import system — making it easy to add new sources like this.

Reflections

The biggest surprise here wasn’t the technical challenges — it was realizing how basic features to export data are missing from major apps in 2025.

Wealthsimple’s Credit Card experience is polished in a lot of ways, but not giving users easy access to their own transaction data feels like a gap. When that happens, as developers, we don't have to wait for permission — we can build the tools we need.

With just a little code and some digging into the page structure, I was able to create a fully working exporter — turning a missing feature into a solved problem in less than half an hour. It could have taken longer, but when you can get ChatGPT to write most of the script, it speeds things up 😄.

This is what I love about being a developer and the power of code:

See a gap
Build a bridge
Keep moving

Tools like this don't just save time — they create freedom as you don't have to rely on someone else deciding what's important for you.

Thanks for reading! Stay tuned for more post about things in my life related technology, my homelab, home improvement, etc!

Appendix

Wealthsimple Credit Card transaction download script

(function() {
  const rows = [];

  const buttons = document.querySelectorAll("button");

  buttons.forEach(button => {
    let date = null;
    let el = button;
    while (el && !date) {
      el = el.previousElementSibling || el.parentElement;
      if (el && el.tagName === "H2" && /\d{4}|Today|Yesterday/i.test(el.textContent)) {
        date = el.textContent.trim();
      }
    }

    const ps = button.querySelectorAll("p");
    const description = ps[0]?.textContent.trim();

    let amount = null;
    ps.forEach(p => {
      if (/\$\d/.test(p.textContent)) {
        amount = p.textContent.trim();
      }
    });

    if (date && description && amount) {
      rows.push([date, description, amount]);
    }
  });

  const uniqueRows = Array.from(new Set(rows.map(e => e.join("|")))).map(r => r.split("|"));

  function csvEscape(value) {
    if (typeof value !== 'string') return value;
    if (value.includes(',') || value.includes('"') || value.includes('\n')) {
      return `"${value.replace(/"/g, '""')}"`; // Escape double quotes by doubling them
    }
    return value;
  }

  const csvHeader = ["Date", "Description", "Amount"];
  const csvRows = [csvHeader, ...uniqueRows];

  const csvContent = "data:text/csv;charset=utf-8,"
    + csvRows.map(row => row.map(csvEscape).join(",")).join("\n");

  const encodedUri = encodeURI(csvContent);
  const link = document.createElement("a");
  link.setAttribute("href", encodedUri);
  link.setAttribute("download", "transactions.csv");
  document.body.appendChild(link);
  link.click();
  document.body.removeChild(link);
})();