Wunder Agent Chat Documentation

Introduction

Wunder Agent Chat is a web-based AI assistant designed for patent attorneys and IP professionals. It provides natural language research capabilities, document analysis, and access to IP databases through an intelligent chat interface. This documentation is intended for IT administrators and technical staff at law firms and IP departments.

Key Features

• AI-powered chat interface for patent research and analysis
• Document upload and extraction (PDF, DOCX, images with OCR)
• Real-time web search integration
• Voice recording and multilingual transcription
• Export functionality (PDF, DOCX, YAML)
• Multi-organization support with email-based authentication
• GDPR-compliant privacy controls

System Requirements

Browser Requirements

Technical Requirements

• JavaScript enabled
• Cookies enabled (required for authentication)
• LocalStorage support
• WebSocket support (for real-time streaming)
• Web Audio API (for voice recording features)
• Stable internet connection (5 Mbps+ recommended)

Network Configuration

Client-Side Requirements

The user's browser only needs to access the Wunder Agent Chat application itself. All external API calls (AI services, search, document processing) are made server-side by the Wunder Agent Chat backend.

Required Domain

Optional Client-Side Domain

The following domain is accessed from the user's browser only if session recording is enabled:

Firewall Considerations

• All communication uses HTTPS (port 443)
• WebSocket connections are used for real-time AI response streaming
• No inbound connections are required from Wunder Agent Chat to client networks

Authentication & Security

Authentication Method

Wunder Agent Chat uses passwordless email authentication via One-Time Passwords (OTP):

1. User enters their email address
2. A 6-digit verification code is sent to that email
3. User enters the code to complete authentication
4. A secure session is created for 7 days

Email Delivery

Verification emails are sent from noreply@wunder-ip.com. Ensure this sender address is not blocked by your email spam filters or security policies.

Session Management

Sessions automatically expire after 7 days of inactivity. Users will need to re-authenticate with a new OTP code.

Security Measures

• No passwords stored: Eliminates password-related vulnerabilities
• Secure cookies: httpOnly flag prevents JavaScript access
• Organization isolation: Users can only access data within their organization
• HTTPS enforced: All traffic encrypted in transit
• Email verification: Required for sensitive account changes

Data & Privacy

Data Collected

User Account Data

• Email address (used as user identifier)
• Organization affiliation
• Theme preference (light/dark mode)
• Tracking preference setting

Usage Data

• Chat conversations and AI responses
• Uploaded document metadata (filename, type, size, page count)
• Tool execution history within chats

Optional Analytics (Opt-in Only)

• Session recordings for quality improvement
• Feature usage patterns
• Error events

Data Storage

Privacy Controls

Users have the following privacy controls available in Settings:

• Usage Tracking: Toggle to enable/disable analytics sharing (disabled by default)
• Account Deletion: Permanently delete all account data and chat history
• Email Change: Update account email with verification

GDPR Compliance

• Privacy by default (tracking opt-in)
• Right to erasure (account deletion)
• Data portability (export features)
• Transparent data collection
• Session recordings sanitize sensitive data automatically

File Upload & Document Processing

Supported File Types

Processing Pipeline

1. User uploads file through the chat interface
2. File is sent to the document processing service
3. Text is extracted (with OCR for images/scanned PDFs)
4. Extracted content is attached to the chat context
5. AI can reference document content in responses

Voice Recording & Transcription

Requirements

• Browser microphone permission must be granted
• Supported audio is recorded directly in the browser
• Transcription supports multiple languages automatically

Supported Languages

The transcription service (whisper-large-v3) supports automatic language detection and transcription for major languages including:

• English, German, French, Spanish, Italian
• Chinese, Japanese, Korean
• And many more

Troubleshooting Voice Features

Export Features

Available Export Formats

Export Content

Exports include:

• Full chat conversation
• AI responses with formatting preserved
• Tool execution results
• Timestamps

Troubleshooting

Common Issues

Authentication Problems

Chat & AI Problems

Document Upload Problems

Browser Compatibility

Diagnostic Steps

If users experience issues:

1. Check browser console (F12 → Console) for JavaScript errors
2. Check network tab (F12 → Network) for failed requests
3. Verify network access to required domains listed above
4. Clear browser cache and try again
5. Try incognito/private mode to rule out extensions
6. Test with another browser to isolate browser-specific issues

Integration Capabilities

Current Integrations

Enterprise Features

Support & Feedback

Getting Help

For technical issues or questions:

1. Consult this documentation first
2. Contact your organization's IT administrator
3. Reach out to Wunder IP support

Reporting Issues

When reporting issues, please include:

• Browser type and version
• Operating system
• Screenshot of any error messages
• Steps to reproduce the issue
• Approximate time the issue occurred

Appendix

Quick Reference: Network Requirements

Required (user browser):
└── chat.wunder-ip.com (HTTPS/443)

Quick Reference: File Limits

Maximum file size: 10 MB
Supported formats: PDF, DOCX, PNG, JPEG, WebP, GIF, TXT, MD

Quick Reference: Session Information

Session duration: 7 days
Cookie type: Secure, httpOnly
Authentication: Email OTP (passwordless)