Session Analysis

How It Works

When someone uses your app, SupaStory records what they do—clicks, scrolls, form inputs, page changes. Then our AI watches the recording and looks for signs of frustration or confusion.

Think of it like having someone review every user session for you, but it happens automatically.

Recording Happens Automatically

Once you add the SDK, sessions are recorded in the background. You don't need to set up event tracking or add code everywhere.

AI Watches Every Session

Our AI reviews each recording. It's trained to spot patterns that indicate problems—like when users click repeatedly on something that doesn't work.

Problems Become "Insights"

When the AI finds an issue, it creates an insight. Each insight explains what's wrong, shows you the session where it happened, and often includes a code fix.

Your users' personal information stays private — Before any session data is sent for AI analysis, SupaStory scrubs personally identifiable information (PII) like names, email addresses, and other personal details. The AI only sees anonymized interaction data, never raw personal information.

What the AI Catches

Signs Users Are Struggling

The AI recognizes when users are having a hard time:

Rage clicks — Clicking the same thing over and over because it's not working
Dead clicks — Clicking on something that looks like a button but isn't
Form struggles — Getting stuck on a form field, typing and deleting repeatedly
Going in circles — Navigating back and forth, clearly looking for something
Giving up — Starting a form or checkout but leaving before finishing

Technical Problems

The AI also catches errors that affect users:

JavaScript errors — When your code throws an exception
Failed API calls — When network requests return errors (4xx, 5xx)
Missing resources — Images, scripts, or stylesheets that fail to load
Console errors — Anything logged to the browser console as an error

Performance Issues

Slow experiences frustrate users too:

Slow page loads — Pages that take too long to become usable
Laggy interactions — Buttons that don't respond immediately
Content jumping — Page elements shifting around after loading
Input delay — Typing in a form feels slow or choppy

How Insights Are Prioritized

Not all problems are equal. The AI ranks insights by:

How many users are affected — A bug hitting 100 users matters more than one hitting 2
How bad it is — A broken checkout is worse than a slightly confusing label
Impact on conversions — Issues blocking purchases or signups get priority
How fixable it is — Clear, actionable problems rank higher

Watching Session Replays

Every insight links to the session where the problem occurred. In the replay, you can see:

Exactly what the user did — Every click, scroll, and keystroke
What was on screen — A visual reconstruction of the page
Network activity — Which API calls succeeded or failed
Errors — Any JavaScript errors that happened

This helps you understand the problem before fixing it.

AI-Generated Fixes

For many issues, the AI doesn't just find the problem—it writes code to fix it.

Here's an example. The AI noticed users rage-clicking a submit button because nothing happened when they clicked:

// The problem: button had no onClick handler
<button className="submit-btn">Submit</button>

// The fix: add the handler and show loading state
<button
  className="submit-btn"
  onClick={handleSubmit}
  disabled={isSubmitting}
>
  {isSubmitting ? 'Submitting...' : 'Submit'}
</button>

The fix shows up as a pull request on GitHub, ready for you to review and merge.

Tips for Better Results

Give the AI context about your app — In your dashboard, tell SupaStory what framework you use, what libraries you have, and how your code is structured. Better context means better fixes.

Check insights regularly — The AI finds issues daily. A quick 5-minute review each morning keeps problems from piling up.

Start with high-impact issues — Focus on insights marked as critical or high severity first. These are the ones hurting users the most.

Use session replays to verify — Before fixing something, watch the replay to make sure you understand the problem.

On this page