AI Features

Virza uses AI to help you understand your research faster. All AI processing respects your privacy. Your documents are never used to train models.

Chat with Virza

Virza’s chat is your intelligent research assistant. You can ask questions about your workspace documents, ask general knowledge questions, or blend both, just like talking to ChatGPT, but with your research library as a superpower.

What you can ask

Document questions: grounded in your uploaded research:

• “What is the main finding of this paper?”
• “Summarize the methodology section”
• “What are the limitations the authors mention?”
• “Compare the approaches across my uploaded papers”

General knowledge questions: answered from Virza’s training knowledge:

• “What is quantum computing?”
• “Is Georgia Tech a good university?”
• “Explain how neural networks work”
• “What are the latest trends in machine learning?”

Hybrid questions: blending your documents with general knowledge:

• “How does my paper’s methodology compare with industry best practices?”
• “What is the significance of these findings in the broader context?”
• “Are the results in my paper consistent with current understanding?”

How it works

Citations and evidence

When Virza answers from your documents, it includes:

• Source cards showing which documents were used, with page numbers
• Inline citations linking claims to specific evidence
• Evidence directions indicating whether sources support, contrast, or mention a topic

You can click any source card to see the full context in the evidence panel.

Tips for better answers

• Be specific: “What methodology does Smith et al. use?” works better than “Tell me about the methodology”
• Use document scope for single-document questions. It narrows the search and improves accuracy
• Use workspace scope for cross-document analysis or when you want Virza to find relevant material
• Ask follow-ups: Virza remembers your conversation context and can build on previous answers

Transparency settings

Control how much detail Virza shows about its reasoning. Open the AI Profile controls (sliders icon in the chat toolbar) and expand the Advanced section to find the Transparency toggle:

• Minimal: Just the knowledge source badge (workspace/general). Clean, distraction-free.
• Standard (default): Adds evidence strength tiers on source cards and warnings when evidence is weak.
• Verbose: Full research mode. Adds a confidence meter showing % of grounded claims, inline color-coded borders distinguishing cited content from general knowledge, and auto-expanded deep research timelines.

Document summaries

Virza automatically generates summaries when you upload a document:

• Executive summary: A concise overview of the paper’s key findings and contributions

On Pro and higher plans, you also get:

• Multi-level summaries: TLDR, structured breakdown (Objective, Methods, Results, Conclusions), and detailed narrative summary
• QA precompute: Pre-generated question-answer pairs so you can quickly quiz a document

Free and Starter plans use a standard AI model for summaries. Pro plans use an advanced model for higher accuracy and depth.

Access summaries from the Summary tab in the document sidebar.

Research intelligence (Pro)

When you have multiple documents in your workspace, Virza can:

• Extraction tables: Build structured comparison tables across your documents with AI-powered extraction
• Compare boards: Side-by-side document comparison with structured diff analysis
• Research intelligence dashboard: AI-powered trends, suggestions, and workspace insights
• Knowledge graph: Visualize citation relationships across your entire library
• Citation graph: See who cites whom, co-citation networks, and citation influence
• Contradiction detection: Spot conflicting findings across papers automatically
• Research gap detection: Find topics your library doesn’t address
• Recommendations: Discover related papers based on shared themes and citations
• Evidence synthesis: Strength scoring and quality assessment across your research
• Discovery feed: Personalized paper recommendations based on your research context

Deep research

For complex questions that require evidence from multiple documents, enable Deep Research from the chat toolbar. Virza will:

1. Analyze gaps in the initial evidence: what’s missing?
2. Generate follow-up queries targeting specific gaps (methodology, statistics, contradictions)
3. Retrieve additional evidence across up to 3 iterative rounds
4. Show the reasoning chain: a collapsible timeline shows each hop, what gap was found, and what was retrieved

Deep research is ideal for literature reviews, systematic comparisons, and questions that span multiple papers.

Export citations (BibTeX), Starter+

After any chat response that cites your documents, click the BibTeX button in the response action bar to export all cited sources as a .bib file. Citation export requires a Starter plan or higher. The export includes:

• Author names, title, journal, year
• DOI with clickable link
• Proper BibTeX formatting for direct use in LaTeX, Overleaf, or reference managers

Evidence quality indicators

Virza gives you transparency about the quality of every response:

• Evidence strength: Each source card shows “Strong”, “Good”, or “Weak” match quality
• Evidence sufficiency: When evidence may not fully address your question, an amber or red badge warns you before you read the answer
• Confidence meter: A small bar below verified responses shows what percentage of claims are grounded in your documents
• Inline provenance: In hybrid responses, cited content has a green left border and general knowledge has an amber left border, so you always know the source

AI in notes

While writing notes, you can invoke AI assistance:

• Paste a quote and ask the AI to explain or contextualize it
• Draft a paragraph summarizing insights from multiple papers
• Generate questions for further research based on your notes

Privacy commitment

• Your documents are processed on secure infrastructure
• Content is never used to train AI models
• AI conversation history is stored securely within your workspace so you can revisit past conversations. Only workspace members with appropriate roles can access it
• Workspace admins can audit AI usage from Settings
• See Privacy for full details on how your data is handled