The chat is Libra’s most versatile feature. You can ask questions, request drafts, analyze documents, and conduct research - all through a natural conversational interface.
Getting Started
Type your message in the input field and press Enter. You can continue the conversation with follow-up questions, and Libra will remember the context from earlier in the chat.Input Options
You can interact with the chat in various ways:| Option | Description |
|---|---|
| Upload documents | Attach files for Libra to analyze or reference |
| Voice input | Click the microphone to dictate instead of typing |
| Research Mode | Access legal databases for sourced research (see Research Mode) |
| Deep Thinking | Extended reasoning for complex questions (see Deep Thinking) |
| Workflows | Run automated multi-step processes (see Workflows) |
| Assistants | Use specialized assistants for specific tasks (see Assistants) |
Working with Responses
Below each response, you can:| Action | Description |
|---|---|
| Copy | Copy with formatting preserved |
| Export | Save as Word, Excel, or email draft |
| Regenerate | Get a different version if needed |
| Send email | Send the response directly via email |
The info banner appears when your response was not generated using expert research sources. To ensure answers are backed by trusted legal sources, enable Research Mode.
Tips for Better Results
Be specific
Be specific
Words like “review,” “check,” or “assess” are often too vague on their own. Review against what standard? Assess which elements? Always specify:
- What to analyze: “Review this NDA for non-compete clauses that may be unenforceable under California law”
- What outcome you need: “Identify any provisions that conflict with GDPR requirements”
- The relevant parties: “Analyze from the perspective of the licensor”
Provide context
Provide context
Context dramatically improves response quality. Include:
- Jurisdiction: Specify the applicable law or court rules (e.g., “under German commercial law” or “per New York state contract law”)
- Background: Explain the business situation or transaction type
- Relevant facts: Include key dates, party relationships, or transaction values
- Purpose: Explain why you need this analysis (e.g., “for due diligence” or “to advise the client on risk”)
Request a format
Request a format
Specify exactly how you want the response structured:
- “Provide your analysis as a numbered list of issues”
- “Create a comparison table with columns for clause, risk level, and recommendation”
- “Draft this as a client-facing memo with an executive summary”
- “Use bullet points for the key findings, then provide detailed analysis below”
Iterate
Iterate
Start broad, then narrow down through follow-up questions. This iterative approach lets you:
- Build context cumulatively: Each exchange adds to Libra’s understanding
- Refine focus: “Now focus specifically on Section 4.2 and its interaction with the indemnification clause”
- Request alternatives: “Give me a more formal version” or “Simplify this for a non-lawyer audience”
- Dig deeper: “Explain the reasoning behind your third point”
Verify outputs
Verify outputs
Treat AI outputs like work from a junior lawyer—trust but verify:
- Check citations: If Libra references cases or statutes, verify they exist and say what’s claimed
- Validate reasoning: Ensure the legal logic is sound for your jurisdiction
- Review for completeness: Consider whether any important issues were missed
- Use Research Mode: For legal research tasks, enable Research Mode to get sourced citations from legal databases
Work with client information
Work with client information
Libra is designed with legal data protection requirements in mind. Your data is stored on servers in Germany, encrypted at rest and in transit, and never used to train AI models.When working with client matters, you can confidently use Libra knowing that your inputs and outputs remain private and are not shared with other customers or third parties outside the EU.For details on our GDPR compliance, data processing agreements, and security controls, see Data Protection & GDPR.