AI-Generated Summaries

AI-generated summaries extract key findings from your research sessions, making it easy to review, share, and reference your work.

What Are AI Summaries?

An AI summary is a concise distillation of a research session's conversation history. The AI:

• Extracts key findings, data points, and conclusions
• Organizes information into logical sections
• Highlights important metrics and comparisons
• Removes conversational filler and redundancy
• Provides a standalone document that captures the essence of your research

Why Use Summaries?

Quick Review

• Save Time: Review hours of research in minutes
• Refresh Memory: Quickly recall findings from old sessions
• Track Progress: See how your understanding evolved

Better Sharing

• Clean Presentation: Share polished summaries instead of raw chat logs
• Focus on Insights: Highlight conclusions rather than the research process
• Professional Format: Present findings to teams or governance forums

Research Organization

• Create References: Build a library of summarized research
• Compare Sessions: Easily compare findings across multiple sessions
• Export Knowledge: Use summaries as the basis for reports or proposals

Generating a Summary

Summary Structure

AI-generated summaries follow a consistent structure:

1. Executive Summary

Brief overview (2-3 sentences) answering:

• What was researched?
• What were the main findings?
• What's the key takeaway?

Example:

This research analyzed yield opportunities for USDC on Base network,
comparing Morpho, Aave, and Aerodrome strategies. Morpho offers the
best risk-adjusted returns at 4.2% APY with strong security. Aerodrome
LP positions offer higher yields (8.5%) but introduce impermanent loss risk.

2. Research Questions

The main questions that guided the research.

Example:

- What are the current USDC yield options on Base?
- How do risk profiles compare across protocols?
- Which protocols have recent security audits?

3. Key Findings

Bulleted list of main discoveries and insights.

Example:

- Morpho-Seamless offers 4.2% APY on USDC with very low risk
- Aave V3 provides 3.8% APY, slightly lower due to different optimization
- Aerodrome USDC-USDT LP yields 8.5% but carries IL risk
- All three protocols have recent audits (2024) and active bug bounties
- Morpho shows highest GitHub activity (commits this week)

4. Data Summary

Table or structured presentation of key metrics.

Example:

| Protocol | APY | Risk | TVL | Last Audit |
|----------|-----|------|-----|------------|
| Morpho   | 4.2%| Low  | $2.3M | Q2 2024  |
| Aave V3  | 3.8%| Low  | $12M  | Q1 2024  |
| Aerodrome| 8.5%| Med  | $45M  | Q2 2024  |

5. Analysis & Conclusions

Interpretation of findings and recommendations.

Example:

For conservative USDC yield on Base, Morpho provides the optimal
risk-adjusted return. While Aerodrome offers 2x higher yields, the
impermanent loss risk makes it suitable only for risk-tolerant investors.
Aave is the most battle-tested but offers slightly lower yields.

Recommendation: Start with Morpho for core holdings, consider Aerodrome
for a smaller allocation if comfortable with LP mechanics.

6. Follow-Up Questions

Areas for deeper investigation.

Example:

- How does Morpho's APY trend over 90 days?
- What percentage of Aerodrome LP returns come from fees vs. emissions?
- Are there alternative LP strategies with lower IL risk?

Customizing Summaries

Custom Prompts

Provide specific instructions when generating summaries:

Generate a summary of this research session.

Summarize this session with emphasis on security risks and potential issues.

Create a technical summary highlighting on-chain metrics and contract details.

Generate a brief executive summary suitable for non-technical audiences.

Summarize this session focusing on the comparison between protocols.

Selective Summarization

Generate summaries for specific parts of a session:

Best Practices

Timing

Generate summaries when:

• ✅ You've completed a research objective
• ✅ The session has 10+ substantive message exchanges
• ✅ You've received data from multiple sources
• ✅ You're ready to share findings

Don't generate too early:

• ❌ After just 2-3 questions
• ❌ Before receiving data-driven responses
• ❌ When still exploring initial questions

Quality Tips

For better summaries:

1. Ask clear questions during research
2. Request specific data rather than vague queries
3. Confirm findings with follow-ups
4. State conclusions explicitly in the chat
5. Highlight important insights by referencing them

Example of explicit conclusion:

Based on this analysis, Morpho appears to be the best option for
conservative USDC yield on Base due to its risk-adjusted returns
and strong security profile.

This helps the AI identify your key takeaways.

Editing Summaries

Always review AI-generated summaries:

• Check accuracy of data points
• Add context that may be missing
• Clarify ambiguities in the AI's interpretation
• Add your own conclusions if the AI missed them
• Update if research continues after summary generation

Sharing Summaries

Summaries are automatically included when you share a research session:

Community members can:

• Read the summary without opening the full session
• Click through to see the detailed conversation
• Use your summary as a starting point for their own research

Exporting Summaries

Export summaries in multiple formats:

Click "Export" → "Markdown (.md)"

Click "Export" → "PDF"

Click "Export" → "JSON"

API Access

Generate summaries programmatically:

POST /api/research/generate-summary
{
  "sessionId": "session_abc123",
  "customPrompt": "Focus on security findings",
  "messageIds": ["msg_1", "msg_2", ...] // Optional: specific messages
}

Response:

{
  "summary": "...",
  "generatedAt": "2025-10-03T12:00:00Z",
  "messageCount": 24,
  "wordCount": 450
}

See API Reference for complete documentation.

Examples

Protocol Comparison Summary

Token Analysis Summary

Troubleshooting

Next Steps