Frequently Asked Questions

This document provides answers to common questions about Qwello, organized by category.

General Questions

What is Qwello?

Qwello is an advanced knowledge graph platform that extracts structured information from unstructured documents. It uses AI to identify entities, relationships, and concepts within documents and organizes them into an interactive knowledge graph. This enables users to discover connections, explore complex information, and gain deeper insights from their documents.

How does Qwello work?

Qwello works through a multi-step process:

1. Web Search: Qwello extracts related documents via the web.

2. Document Upload: Users upload PDF documents to the platform.

3. Text Extraction: Qwello extracts text content from the documents.

4. Entity Recognition: AI models identify important entities (people, organizations, concepts, etc.) in the text.

5. Relationship Extraction: The system identifies relationships between entities.

6. Knowledge Graph Construction: Entities and relationships are organized into a structured knowledge graph.

7. Visualization: The knowledge graph is presented in an interactive visualization.

8. Search and Analysis: Users can search, filter, and analyze the knowledge graph to discover insights.

What types of documents can Qwello process?

Qwello primarily processes PDF documents, including:

• Research papers and academic articles

• Business reports and white papers

• Legal documents and contracts

• Technical documentation

• Patents and intellectual property documents

• News articles and publications

• Financial reports and analysis

The system works best with text-based PDFs rather than scanned documents, although it can process scanned documents with OCR (Optical Character Recognition) capabilities.

What are the main benefits of using Qwello?

The main benefits of using Qwello include:

• Time Savings: Quickly extract key information from large documents without manual reading.

• Knowledge Discovery: Uncover hidden connections and relationships across multiple documents.

• Enhanced Understanding: Visualize complex information in an intuitive, interactive format.

• Improved Decision Making: Make more informed decisions based on comprehensive information analysis.

• Knowledge Sharing: Easily share insights and findings with colleagues and stakeholders.

• Research Acceleration: Speed up research processes by quickly identifying relevant information.

• Institutional Knowledge: Preserve and leverage organizational knowledge more effectively.

Who is Qwello designed for?

Qwello is designed for knowledge workers and organizations that deal with large volumes of complex information, including:

• Researchers and Academics: For literature review, research synthesis, and knowledge discovery.

• Business Analysts: For market research, competitive analysis, and strategic planning.

• Legal Professionals: For contract analysis, case research, and regulatory compliance.

• Healthcare Professionals: For medical literature review, clinical research, and patient data analysis.

• Government Agencies: For policy analysis, intelligence gathering, and information management.

• Financial Analysts: For financial report analysis, risk assessment, and investment research.

• Knowledge Management Teams: For organizing and leveraging institutional knowledge.

Technical Questions

What technologies does Qwello use?

Qwello is built on a modern technology stack that includes:

• Frontend: React, TypeScript, and D3.js for visualization

• Backend: NestJS (Node.js framework) with TypeScript

• Database: MongoDB for document and knowledge graph storage

• Caching: Redis for performance optimization

• AI Processing: Integration with advanced AI models for entity and relationship extraction

• Queue Management: Bull for processing job queues

• Authentication: JWT-based authentication system

• Containerization: Docker for deployment and scaling

• Orchestration: Kubernetes for production environments

How does Qwello handle large documents?

Qwello employs several strategies to handle large documents efficiently:

1. Chunking: Large documents are broken down into manageable chunks for processing.

2. Streaming Processing: Documents are processed in a streaming fashion rather than loading the entire document into memory.

3. Distributed Processing: Processing tasks are distributed across multiple worker instances.

4. Asynchronous Processing: Documents are processed asynchronously in the background, allowing users to continue working while processing occurs.

5. Progressive Loading: Results are loaded progressively as they become available.

6. Optimized Storage: Processed results are stored in an optimized format for quick retrieval.

What entity types does Qwello recognize?

Qwello recognizes a wide range of entity types, including:

• People: Individuals mentioned in the document

• Organizations: Companies, institutions, agencies, etc.

• Locations: Geographic places, countries, cities, etc.

• Concepts: Abstract ideas, theories, methodologies

• Events: Occurrences with a specific time and place

• Products: Commercial offerings, services, etc.

• Technologies: Technical systems, software, hardware, etc.

• Documents: References to other documents or publications

• Time Periods: Specific dates, time ranges, eras

• Measurements: Quantities, statistics, metrics

Additionally, Qwello supports custom entity types that can be defined based on specific domain needs.

What relationship types does Qwello identify?

Qwello identifies various relationship types between entities, including:

• Association: General connection between entities

• Causation: One entity causes or influences another

• Part-Whole: One entity is part of another

• Temporal: Time-based relationships (before, after, during)

• Spatial: Location-based relationships

• Hierarchical: Parent-child or superior-subordinate relationships

• Functional: One entity serves a function for another

• Comparative: Similarities or differences between entities

• Transformational: One entity transforms into another

• Attribution: Properties or characteristics of entities

Custom relationship types can also be defined for specific domains or use cases.

How accurate is Qwello's entity and relationship extraction?

Qwello's extraction accuracy varies depending on several factors:

• Document Quality: Clearer, well-structured documents yield better results

• Domain Specificity: Performance is better in domains the system has been trained on

• Entity Type: Some entity types (like people and organizations) typically have higher accuracy than abstract concepts

• Relationship Complexity: Simple, explicit relationships are identified more accurately than complex or implicit ones

On average, Qwello achieves:

• 85-95% accuracy for common entity types in well-structured documents

• 75-85% accuracy for relationship extraction between clearly related entities

• Each extracted entity and relationship includes a confidence score, allowing users to filter based on confidence thresholds

Qwello continuously improves its accuracy through model updates and learning from user feedback.

How does Qwello handle data security and privacy?

Qwello implements comprehensive security and privacy measures:

• Data Encryption: All data is encrypted both in transit (TLS) and at rest (AES-256)

• Access Controls: Role-based access control (RBAC) for fine-grained permissions

• Authentication: Secure authentication with MFA support

• Data Isolation: Multi-tenant architecture with strict data isolation

• Compliance: Designed to meet GDPR, CCPA, and other privacy regulations

Deployment and Installation

What deployment options are available for Qwello?

Qwello offers several deployment options:

1. Cloud SaaS: Fully managed cloud service with subscription pricing

2. Private Cloud: Dedicated cloud instance in your preferred cloud provider

3. On-Premises: Self-hosted deployment within your own infrastructure

4. Hybrid: Combination of cloud and on-premises components

The deployment option you choose depends on your security requirements, existing infrastructure, and organizational policies.

What are the system requirements for on-premises deployment?

For on-premises deployment, the minimum system requirements are:

Production Environment (Recommended):

• Kubernetes Cluster: Minimum 3 nodes

• Node Specifications: 8 CPU cores, 32GB RAM per node

• Storage: 100GB+ SSD storage for the database

• Operating System: Linux (Ubuntu 20.04 LTS or later recommended)

• Network: Outbound internet access for AI API calls (if using cloud AI services)

Minimal Development Environment:

• Single Server: 4 CPU cores, 16GB RAM

• Storage: 50GB SSD storage

• Operating System: Linux, macOS, or Windows with Docker support

• Docker: Docker and Docker Compose

Usage and Features

How do I upload documents to Qwello?

Documents can be uploaded to Qwello through several methods:

1. Web Interface: Drag and drop files into the upload area or use the file picker

2. Batch Upload: Upload multiple files at once through the web interface

3. API: Programmatically upload documents using the REST API

4. Folder Monitoring: Set up a monitored folder where new documents are automatically processed

5. Email Integration: Send documents as email attachments to a dedicated address

6. Cloud Storage Integration: Connect to cloud storage providers (Google Drive, Dropbox, etc.)

How long does it take to process a document?

Processing time depends on several factors:

• Document Size: Larger documents take longer to process

• Document Complexity: Documents with complex content and many entities take longer

• System Load: Processing may be slower during peak usage times

• Processing Configuration: More thorough processing settings increase processing time

Typical processing times:

• Small documents (1-10 pages): 30 seconds to 2 minutes

• Medium documents (11-50 pages): 2-5 minutes

• Large documents (51-200 pages): 5-15 minutes

• Very large documents (200+ pages): 15+ minutes

Processing happens asynchronously, so you can continue using the system while documents are being processed.

How do I search within the knowledge graph?

Qwello offers several ways to search and explore the knowledge graph:

1. Text Search: Search for entities and relationships by name or content

2. Advanced Search: Use filters for entity types, relationship types, confidence scores, etc.

3. Query Language: Use the graph query language for complex searches

4. Visual Exploration: Navigate the graph visually by expanding nodes and following relationships

5. Saved Searches: Save and reuse common searches

6. Search Templates: Use pre-defined search templates for common use cases

Can I export data from Qwello?

Yes, Qwello supports various export formats:

• JSON: Complete knowledge graph data in JSON format

• Markdown: Download document data in Markdown format

• Reports: Formatted reports with visualizations and analysis

How does Qwello handle multiple documents?

Qwello can process multiple documents in several ways:

1. Individual Processing: Each document is processed separately with its own knowledge graph

2. Batch Processing: Multiple documents are processed as a batch

3. Merged Knowledge Graphs: Knowledge graphs from multiple documents can be merged

4. Cross-Document Analysis: Identify connections and relationships across documents

5. Document Collections: Organize documents into collections for joint analysis

6. Comparative Analysis: Compare knowledge graphs from different documents

Pricing and Licensing

What pricing models are available?

Qwello offers several pricing models:

1. Free Tier: Limited functionality for individual users (processing limits apply)

2. Professional: Per-user subscription with standard features

Visit qwello.ai/pricing for current pricing details.

What additional features are available in paid tiers?

Paid tiers offer additional features such as:

• Unlimited document processing (subject to fair use policy)

• Larger document support (up to 1000+ pages)

• Priority support

Support and Resources

How can I get help with Qwello?

Qwello offers several support channels:

1. Documentation: Comprehensive documentation at docs.qwello.ai
2. Community Forum: User community for questions and discussions

3. Email Support: Email support for all paid plans

How do I report bugs or request features?

You can report bugs or request features through several channels:

1. Email: Send reports to [e-mail]
2. GitHub: Submit issues on GitHub

Is there a roadmap for future features?

Yes, Qwello maintains a public roadmap that outlines planned features and improvements. The roadmap is available at qwello.ai/roadmap and is updated quarterly.

Major upcoming features include:

• Enhanced multilingual support

• Advanced document comparison

• Expanded visualization options

• Additional integration capabilities

• Mobile application

• Collaborative annotation features

• Custom AI model training interface

Enterprise customers can also influence the roadmap through the customer advisory board.

Technical Support

What browsers are supported?

Qwello supports the following browsers:

• Chrome: Version 90 and later

• Firefox: Version 88 and later

• Safari: Version 14 and later

• Edge: Version 90 and later (Chromium-based)

Mobile browsers are supported on a best-effort basis, but the full desktop experience is recommended for optimal usage.

Does Qwello work on mobile devices?

Qwello offers varying levels of mobile support:

• Responsive Web Interface: The web interface adapts to smaller screens, though some advanced features may be limited

• Mobile Browsers: Supported on recent versions of mobile browsers

For the best experience, especially with complex knowledge graphs, a desktop or laptop computer is recommended.

What languages does Qwello support?

Qwello currently supports the following languages:

User Interface Languages:

• English

• Spanish

• French

• German

• Japanese

• Chinese (Simplified)

Document Processing Languages:

• English (full support)

• Spanish, French, German (good support)

• Italian, Portuguese, Dutch, Russian (moderate support)

• Other languages (basic support)

Language support is continuously expanding, with additional languages planned for future releases.

How do I recover a lost password?

To recover a lost password:

1. Click the "Forgot Password" link on the login page

2. Enter your email address

3. Check your email for a password reset link

4. Click the link and follow instructions to create a new password

How do I contact customer support?

You can contact customer support through several channels:

• Email: [e-mail]

When contacting support, please include:

• Your account information

• Detailed description of the issue

• Steps to reproduce the problem

• Error messages or screenshots

• Browser and operating system information

Additional Questions

Can Qwello be used for languages other than English?

Yes, Qwello supports multiple languages, though with varying levels of capability:

• English: Full support with highest accuracy

• Major European Languages (Spanish, French, German): Good support with strong accuracy

• Other Languages: Basic to moderate support depending on the language

The system is continuously improving its multilingual capabilities. For enterprise customers, custom language models can be developed for specific language needs.

Is Qwello GDPR compliant?

Yes, Qwello is designed with GDPR compliance in mind:

• Data Processing Agreements: Available for all customers

• Data Subject Rights: Tools to fulfill data subject requests

• Data Minimization: Only necessary data is collected and processed

• Consent Management: Features for managing user consent

• Data Protection Impact Assessment: Documentation available for enterprise customers

• Data Breach Notification: Procedures in place for timely notification

• Data Protection Officer: Appointed DPO for GDPR matters

• International Data Transfers: Compliant mechanisms for data transfers

For detailed information on GDPR compliance, please contact [e-mail].

What AI models does Qwello use?

Qwello uses a combination of AI models:

• Large Language Models: For text understanding and knowledge extraction

• Computer Vision Models: For document layout analysis and image processing

• Graph Neural Networks: For relationship inference and knowledge graph enhancement

• Custom Models: Domain-specific models for specialized use cases

The specific models used are regularly updated to incorporate the latest advancements in AI research. Enterprise customers can also integrate custom models for specific domains or use cases.

This FAQ covers the most common questions about Qwello. If you have additional questions, please contact [e-mail] or visit our community forum.