Skip to main content

​
What Is Test Detection?

Detection scans your application to discover testable user flows — like “User can sign up”, “Admin can create a report”, or “Customer can complete checkout”. The agent identifies:
  • Key user journeys through your application
  • Critical business flows
  • Common interaction patterns
  • Edge cases and error scenarios
The output is a set of test flow definitions that describe what should be tested. These flow definitions then feed into test generation where actual Playwright code is created.
Test Detection - User Flows

​
Connecting Your Codebase

If you’ve connected your source code repository in Git Integration, Checksum can analyze your actual source code (routes, components, interactions) during detection. This significantly improves detection accuracy and speed compared to detection without source code access. The detection approach is selected automatically based on whether a code repo is connected — there’s nothing extra to configure.

​
Checksum SDK — Learn from Real Users

For even better detection results, integrate the Checksum JavaScript SDK into your application. The SDK records real user sessions non-intrusively — capturing clicks, navigation, and input patterns — so the agent understands how your app is actually used in production. This real-world behavior data feeds into both detection and generation, helping the agent identify the most important flows and create more accurate tests.
The SDK is optional but recommended for complex applications with nuanced workflows. See the SDK integration guide for installation instructions.

​
Deep vs Standard Detection

Detection supports two modes that determine how thorough the analysis is:

​
Standard

  • Faster, more direct detection
  • Best for well-structured apps where flows are straightforward
  • Gets results quickly

​
Deep

  • Includes a planning and interview phase before detection
  • The agent asks clarifying questions about your app to ensure thorough coverage
  • Better for complex apps with nuanced workflows
  • Produces more comprehensive test flow definitions
See Deep vs Standard for more details on how these modes differ across all Checksum operations.

​
Running Detection

​
From the Web App

  1. Navigate to Test Generation in the sidebar
  2. Select or create a collection (e.g., “Checkout Flows”)
  3. Click the Auto-detect test flows button on the collection
  4. The agent session begins — you can monitor progress and answer any questions
Test Detection - User Flows
Detection creates an agent session that runs autonomously. The agent may pause to ask you questions about your application (e.g., “Which features should I focus on?”, “Are there any areas to exclude?”).

​
From a Coding Agent

You can also run detection from your IDE using the /checksum_detect_tests command. See Coding Agent Integration for details.

​
Reviewing Detected Flows

Once detection completes, the discovered test flows appear in your collection.
Test Detection - User Flows
Each detected flow includes:
  • A title describing the user journey
  • A description of the steps involved
  • The start URL for the flow
You can:
  • Edit any detected flow to refine the description or steps
  • Delete flows that aren’t relevant
  • Add manual flows for scenarios the agent didn’t detect
  • Generate tests from any flow (see Generate Tests)

​
Detection vs Manual Flow Creation

ApproachWhen to Use
Auto-detectionYou want Checksum to discover flows from your app automatically
Manual creationYou have specific test scenarios in mind, or want to test something detection missed
Both approaches produce the same result — test flow definitions that can be used for generation. You can mix and match: detect flows automatically, then add a few manual flows for edge cases.

​
Next Steps