What Happens When a Test Starts?

When you start an exploration, Rihario launches a browser, loads your page, analyzes it with AI, then begins exploring. The process includes diagnosis, understanding page structure, planning exploration, and executing actions. Here's what happens step by step.

Phase 1: Initialization (0-5 seconds)

1. Browser Launch

Rihario launches a fresh browser instance:

• Chrome, Firefox, or Safari (depending on your selection)
• Clean profile (no cookies, cache, or extensions)
• Desktop or mobile viewport (as specified)
• Network conditions configured

2. Page Load

The browser loads your URL:

• Navigates to the specified URL
• Waits for page to load (DOMContentLoaded)
• Waits for network idle (no pending requests)
• Captures initial screenshot

3. Pre-Test Diagnosis

Before exploring, the AI performs a quick diagnosis:

• Analyzes page structure - Understands what's on the page
• Identifies testable elements - Finds buttons, forms, links
• Detects blockers - Checks for CAPTCHAs, cookie banners, auth prompts
• Assesses complexity - Estimates how much exploration is needed

See Pre-Test Diagnosis for details on this phase.

Phase 2: Understanding (5-15 seconds)

1. Page Analysis

The AI deeply analyzes the page:

• Computer vision - "Sees" what the page looks like
• DOM analysis - Reads HTML structure
• Element identification - Finds interactive elements
• Context understanding - Determines page purpose and key elements

2. Exploration Planning

Based on analysis, the AI plans what to explore:

• Prioritizes actions - Decides what's most important
• Follows your instructions - If you provided guidance, focuses on that
• Identifies flows - Recognizes common user flows (login, checkout, etc.)
• Plans sequence - Determines order of actions

3. Safety Checks

Before proceeding, safety checks run:

• Cookie banner detection - Checks if cookie consent is needed
• CAPTCHA detection - Identifies CAPTCHAs that would block exploration
• Auth prompt detection - Recognizes login prompts
• Redirect loop detection - Checks for infinite redirects

Phase 3: Exploration (Ongoing)

1. Action Execution

The AI begins executing actions:

• Clicks buttons - Interacts with elements
• Fills forms - Types in input fields
• Navigates - Follows links to new pages
• Waits - Waits for pages to load, animations to finish
• Scrolls - Scrolls to see content below fold

2. Issue Detection

While exploring, the AI constantly watches for problems:

• Visual issues - Broken layouts, overlapping elements
• Console errors - JavaScript errors logged
• Network errors - Failed requests, timeouts
• Accessibility issues - Missing labels, contrast problems

3. Evidence Collection

When issues are found, evidence is captured:

• Screenshots - Visual proof of problems
• Console logs - JavaScript error logs
• Network logs - Failed request details
• DOM snapshots - HTML structure at time of issue

4. Adaptive Exploration

The AI adapts based on what it finds:

• Adjusts strategy - Changes approach if initial plan doesn't work
• Handles errors - Retries failed actions with different methods
• Explores new paths - Discovers and follows new navigation options
• Stops when stuck - Marks exploration as BLOCKED if it can't proceed

Phase 4: Completion

1. Exploration Finishes

Exploration ends when:

• Natural completion - AI finishes planned exploration
• Time limit reached - Maximum exploration time exceeded
• Blocked - Hit a CAPTCHA, MFA, or other blocker
• Error - Fatal error occurred
• Manual stop - You manually cancelled

2. Result Compilation

Final results are compiled:

• Issues aggregated - All findings grouped by type
• Steps logged - Full sequence of actions recorded
• Evidence organized - Screenshots and logs attached to issues
• Summary generated - Overview of what was explored and found

3. Results Displayed

You see the results:

• Summary - Pages explored, issues found, duration
• Step log - Every action taken
• Issues list - Problems detected with evidence
• Video replay - Watch the entire exploration

Timeline Example

What You See During Execution

Live View

While the exploration runs, you see:

• Current action - What the AI is doing right now
• Step counter - How many steps completed
• Browser view - Screenshot of current page state
• Issues detected - Problems found so far
• Progress indicator - Estimated completion

You Can Intervene

At any time, you can:

• Pause - Stop the exploration
• Take Control - Manually interact, then resume
• Provide Guidance - Give the AI new instructions
• Cancel - Stop completely

Common Scenarios

Scenario 1: Smooth Exploration

Everything works as expected:

1. Page loads quickly
2. AI finds interactive elements easily
3. Exploration proceeds smoothly
4. No blockers encountered
5. Completes successfully

Scenario 2: Cookie Banner

Cookie banner appears:

1. Page loads
2. Cookie banner detected
3. AI attempts to dismiss it
4. If successful: continues exploring
5. If fails: exploration marked as BLOCKED

See How Cookie Banners Are Handled for details.

Scenario 3: CAPTCHA Block

CAPTCHA appears:

1. Page loads
2. CAPTCHA detected
3. AI cannot proceed
4. Exploration marked as BLOCKED
5. You can take control to solve it manually

Next Steps

• Learn about pre-test diagnosis
• See how cookie banners are handled
• Understand skipped and blocked steps