> ## Documentation Index
> Fetch the complete documentation index at: https://onetest.ai/llms.txt
> Use this file to discover all available pages before exploring further.

# AI Assistant Overview

> Your intelligent testing companion powered by Claude

The OneTest AI Assistant is your intelligent testing companion that helps you create tests, search your test suite, analyze results, and get instant answers—all through natural conversation.

<Note>
  **Prerequisites:** The AI Assistant requires an LLM provider to be configured before use. Go to [Settings > LLM Configuration](/settings/llm-configuration) to set up your AI provider. AI capabilities are built into the platform and available across test generation, search, and analysis once your provider is configured.
</Note>

## What Can the AI Do?

<CardGroup cols={2}>
  <Card title="Generate Tests" icon="wand-magic-sparkles" href="/ai/generate-tests">
    Create complete, detailed test cases in seconds by describing what you want to test
  </Card>

  <Card title="Search Tests" icon="magnifying-glass" href="/ai/search-tests">
    Find tests using natural language instead of complex filters
  </Card>

  <Card title="Ask Questions" icon="circle-question" href="/ai/ask-questions">
    Get instant answers about your tests, runs, and quality metrics
  </Card>

  <Card title="Best Practices" icon="lightbulb" href="/ai/best-practices">
    Learn how to get the most out of the AI Assistant
  </Card>
</CardGroup>

## How to Access the AI

<Steps>
  <Step title="Open the Chat">
    Press `Cmd + K` (Mac) or `Ctrl + K` (Windows/Linux)
  </Step>

  <Step title="Type Your Request">
    Describe what you want in plain English
  </Step>

  <Step title="Review Results">
    The AI will generate tests, search results, or answers
  </Step>

  <Step title="Take Action">
    Save, refine, or continue the conversation
  </Step>
</Steps>

## Common Use Cases

<Tabs>
  <Tab title="Generate Tests">
    **Create a single test:**

    ```
    Create a test for login with valid credentials
    ```

    **Create multiple tests:**

    ```
    Generate smoke tests for the checkout flow
    ```

    **From requirements:**

    ```
    Create tests for this user story: [paste story]
    ```

    **From documentation:**

    ```
    Generate tests from these API docs: [paste or upload]
    ```

    <Tip>
      The AI generates complete tests with:

      * Clear titles
      * Step-by-step instructions
      * Expected results
      * Preconditions
      * Tags
    </Tip>
  </Tab>

  <Tab title="Search Tests">
    **By status:**

    ```
    Show me all failed tests from yesterday
    ```

    **By feature:**

    ```
    Find all smoke tests for login
    ```

    **By metadata:**

    ```
    Which tests haven't been run in the last 30 days?
    ```

    **Complex queries:**

    ```
    Show P1 tests that failed in staging but passed in dev
    ```

    <Tip>
      The AI translates your natural language into precise OQL queries automatically.
    </Tip>
  </Tab>

  <Tab title="Ask Questions">
    **About quality:**

    ```
    What's the pass rate for checkout tests this sprint?
    ```

    **About coverage:**

    ```
    Which features don't have any smoke tests?
    ```

    **About trends:**

    ```
    Are we seeing more failures than usual?
    ```

    **About specific tests:**

    ```
    Why is TC-0042 failing?
    ```

    <Tip>
      The AI has access to all your test data and can provide instant insights.
    </Tip>
  </Tab>

  <Tab title="Get Help">
    **Learn features:**

    ```
    How do I organize tests with folders?
    ```

    **Understand concepts:**

    ```
    What's the difference between a test case and a test run?
    ```

    **Troubleshoot:**

    ```
    Why can't I delete this test?
    ```

    **Get suggestions:**

    ```
    What tests should I create for a new feature?
    ```
  </Tab>
</Tabs>

## Example Conversations

<AccordionGroup>
  <Accordion title="Creating a Login Test" icon="user-lock">
    **You:** Create a test for login with valid credentials

    **AI:** I'll create a comprehensive login test for you.

    **Generated Test:**

    ```
    TC-0042: Verify successful login with valid credentials

    Type: Manual
    Priority: P2
    Tags: smoke, login, authentication

    Preconditions:
    - User has a valid account in the system
    - User is not currently logged in
    - Login page is accessible

    Test Steps:
    1. Navigate to the login page
       Expected: Login form displays with username and password fields

    2. Enter valid username: test@example.com
       Expected: Username field accepts input

    3. Enter valid password
       Expected: Password is masked (•••)

    4. Click "Login" button
       Expected: User redirected to dashboard within 2 seconds

    5. Verify user session
       Expected: User name displayed in header: "Welcome, Test User"

    Pass Criteria:
    - Login succeeds without errors
    - Dashboard loads within 2 seconds
    - Session is created and persists
    ```

    **AI:** Would you like me to create additional test cases for invalid credentials or edge cases?
  </Accordion>

  <Accordion title="Finding Failed Tests" icon="triangle-exclamation">
    **You:** Show me tests that failed in the last production run

    **AI:** I found 3 tests that failed in the most recent production run:

    1. **TC-0156: Checkout with saved payment method**
       * Failed at: Step 4 (Payment processing)
       * Error: Payment gateway timeout
       * Last passed: 2 days ago

    2. **TC-0089: User profile update**
       * Failed at: Step 3 (Save changes)
       * Error: 500 Internal Server Error
       * Last passed: 1 week ago

    3. **TC-0234: Search with special characters**
       * Failed at: Step 2 (Execute search)
       * Error: Invalid query format
       * First failure

    Would you like me to:

    * Create bug reports for these failures?
    * Show the full test execution details?
    * Compare with staging results?
  </Accordion>

  <Accordion title="Analyzing Test Coverage" icon="chart-pie">
    **You:** Do we have enough smoke tests for the authentication feature?

    **AI:** Let me analyze your authentication test coverage:

    **Current Coverage:**

    * ✅ Login with valid credentials (TC-0042)
    * ✅ Logout (TC-0043)
    * ✅ Session timeout (TC-0198)
    * ⚠️ Login with invalid credentials (missing)
    * ⚠️ Password reset flow (TC-0167 exists but not tagged 'smoke')
    * ⚠️ "Remember me" functionality (missing)
    * ⚠️ Account lockout after failed attempts (missing)

    **Recommendation:**
    You have 3 smoke tests for authentication, but I recommend adding 4 more to cover critical failure scenarios and security features.

    Would you like me to:

    1. Generate the missing smoke tests?
    2. Update TC-0167 to include 'smoke' tag?
    3. Create a complete authentication test suite?
  </Accordion>
</AccordionGroup>

## Interactive Forms

When creating structured content like test cases, the AI presents **interactive forms** instead of plain text responses. These forms appear as editable cards in the chat:

* **Pre-filled fields** based on your request (title, steps, priority, tags)
* **Edit before saving** — modify any field directly in the form
* **Expand to full view** for complex forms with many fields
* **Submit or cancel** — submit saves the data, cancel dismisses the form

This means you can say "create a test for login" and get a structured, editable form that creates a real test case when you submit it — no copy-pasting needed.

## AI Capabilities

The AI Assistant can:

<CardGroup cols={3}>
  <Card title="Understand Context" icon="brain">
    Knows what page you're on and what you're working with
  </Card>

  <Card title="Access Your Data" icon="database">
    Can see all your tests, runs, and results
  </Card>

  <Card title="Learn Patterns" icon="chart-network">
    Recognizes your team's testing patterns and conventions
  </Card>

  <Card title="Generate Content" icon="file-pen">
    Creates tests, tags, and documentation
  </Card>

  <Card title="Analyze Trends" icon="chart-line">
    Identifies quality trends and potential issues
  </Card>

  <Card title="Provide Insights" icon="lightbulb">
    Offers recommendations based on best practices
  </Card>
</CardGroup>

## Tips for Better Results

<Note>
  **Be specific:** Instead of "create a test," try "create a test for user registration with email verification"
</Note>

<Tip>
  **Provide context:** Mention the feature, user story, or requirement you're testing
</Tip>

<Tip>
  **Ask follow-ups:** The AI remembers your conversation, so you can refine and iterate
</Tip>

<Tip>
  **Upload documents:** You can paste or upload requirements, user stories, or API docs
</Tip>

<Warning>
  **Review AI output:** Always review generated tests to ensure they match your specific requirements
</Warning>

## Keyboard Shortcuts

| Shortcut           | Action            |
| ------------------ | ----------------- |
| `Cmd/Ctrl + K`     | Open AI chat      |
| `Esc`              | Close chat        |
| `↑`                | Edit last message |
| `Cmd/Ctrl + Enter` | Send message      |

## What's Next?

<CardGroup cols={2}>
  <Card title="Generate Your First Test" icon="wand-magic-sparkles" href="/ai/generate-tests">
    Learn how to create tests with AI
  </Card>

  <Card title="Search Like a Pro" icon="magnifying-glass" href="/ai/search-tests">
    Master natural language search
  </Card>

  <Card title="Ask Better Questions" icon="circle-question" href="/ai/ask-questions">
    Get more insights from the AI
  </Card>

  <Card title="Best Practices" icon="star" href="/ai/best-practices">
    Advanced tips and techniques
  </Card>
</CardGroup>