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

# AI Chat

> Get AI assistance understanding and fixing security issues

AI Chat provides an intelligent assistant to help you understand security vulnerabilities, get remediation guidance, and answer questions about your code.

## Accessing AI Chat

### From an Issue

1. Open any issue in the dashboard
2. Click **"Ask AI"**
3. Chat opens with the issue context preloaded

### From the Header

1. Click the **AI Chat** icon in the project header
2. Chat opens for general security questions
3. No specific issue context

## What AI Chat Can Do

### Explain Vulnerabilities

Ask about any security issue:

* "What is SQL injection?"
* "How can this XSS be exploited?"
* "What's the impact of this vulnerability?"

### Suggest Fixes

Get remediation guidance:

* "How do I fix this?"
* "Show me a secure implementation"
* "What library should I use for sanitization?"

### Analyze Code

Discuss code patterns:

* "Is this input validation sufficient?"
* "What security issues might this function have?"
* "How can I make this more secure?"

### Answer Security Questions

General security guidance:

* "What's the best practice for storing passwords?"
* "How should I handle user authentication?"
* "What security headers should I set?"

## Example Conversations

### Understanding an Issue

```
You: What does this SQL injection issue mean?

AI: This SQL injection vulnerability occurs because user input
is directly concatenated into the SQL query string. An attacker
could input something like `' OR '1'='1` to bypass authentication
or `'; DROP TABLE users; --` to delete data.

The vulnerable code builds the query like:
"SELECT * FROM users WHERE id = " + userId

To fix this, use parameterized queries...
```

### Getting a Fix

```
You: How do I fix this XSS vulnerability?

AI: For this XSS issue where user input is being rendered in HTML,
you have several options:

1. Use textContent instead of innerHTML:
   element.textContent = userInput;

2. Use a sanitization library:
   element.innerHTML = DOMPurify.sanitize(userInput);

3. Encode the output for the context...
```

## Context Awareness

When opened from an issue, AI Chat knows:

* The vulnerability type and description
* The affected code snippet
* The file path and location
* Your codebase language and framework

This context helps provide relevant, specific guidance.

## Tips for Better Responses

### Be Specific

```
❌ "How do I fix this?"
✅ "How do I fix this SQL injection using Prisma ORM?"
```

### Provide Context

```
❌ "Is this secure?"
✅ "Is this secure for a public API endpoint that handles payments?"
```

### Ask Follow-ups

The chat maintains conversation history, so you can:

* Ask clarifying questions
* Request more detail
* Ask for alternative approaches

## Limitations

AI Chat is designed to help, but:

* Always verify suggestions before implementing
* It may not know your specific business logic
* Complex architectural questions may need human review
* It cannot access external systems or run code

## Chat History

* Chat history persists during your session
* History is cleared when you close the browser
* Conversations are not stored permanently

## Related Pages

<CardGroup cols={2}>
  <Card title="Issues" icon="bug" href="/dashboard/issues">
    View and manage findings
  </Card>

  <Card title="Learnings" icon="brain" href="/dashboard/learnings">
    Create learnings from issues
  </Card>

  <Card title="Security Categories" icon="shield" href="/security/overview">
    Learn about vulnerability types
  </Card>

  <Card title="How It Works" icon="lightbulb" href="/how-it-works">
    Understand AI detection
  </Card>
</CardGroup>