Terms of Use

11.1 Scope

This section applies to the LITIGAGENT case workspace tools, including case file upload and analysis, case chat, objection drafting, and discovery tools used within the case workspace. It supplements the general AI limitations described in Section 5 above, which continue to apply to all AI-generated output.

11.2 How We Protect Your Data

We take the following measures to protect case materials processed through the Service:

• Local-first architecture. When you upload case files, we extract the text, run OCR on scanned documents, build a searchable index, and store everything on our own servers. None of these steps involve any third party. Your files never leave our infrastructure.
• Entity obfuscation at the AI boundary. Before we send any text to our AI provider, we automatically scan it for identifying information — party names, attorney names, company names, email addresses, phone numbers, and case numbers — and replace each one with a generic placeholder like PERSON_1 or COMPANY_1. The AI sees "PERSON_1 was employed by COMPANY_1" instead of real names. When the AI responds using those same placeholders, we swap the real names back in before you see the answer.
• High-accuracy seeding from case metadata. When you upload a complaint and we extract the party names, case number, and attorney information, that information becomes a seed list for obfuscation. Because we already know exactly who the parties are, we can find and replace those names with near-perfect accuracy.
• Encryption at rest. Your case files, extracted text, notes, and chat history are encrypted in our database. Even if someone gained unauthorized access to the raw database file, the content would be unreadable without the encryption key.
• No-training guarantee. Anthropic (our AI provider) does not use any data sent through their API to train their AI models. This is a contractual commitment documented in their commercial API terms.
• Time-limited retention by AI provider. Anthropic retains API data for up to 30 days for trust and safety monitoring, then permanently deletes it. During that window, the data they hold is the obfuscated version — placeholders, not real names.

11.3 Limitations

No obfuscation system is perfect, and we do not claim otherwise. You should understand these limitations before using the case workspace tools:

1. Entity detection is imperfect. Our natural language processing may not catch every name, especially unusual names, names that appear only in context (e.g., "the VP of Marketing"), or names embedded in non-standard formatting. Known parties from your case metadata are handled with high accuracy, but other entities not detected by our pattern matching or natural language processing will pass through to the AI provider.
2. Contextual identification is possible. Even with all names replaced, the factual pattern of a case may be identifying. We do not attempt to obfuscate facts, dates, locations, or circumstances — doing so would destroy the legal analysis that makes the tool useful.
3. AI provider data retention. During the 30-day retention window, Anthropic holds obfuscated text. While this text contains placeholders rather than real names, it still contains the factual substance of the case. We consider this risk low (Anthropic is SOC 2 Type II certified), but it is not zero.
4. Encryption protects at the storage level. Our encryption at rest protects against database file theft, but data is necessarily decrypted in application memory during normal use. A compromise of the application server would expose decrypted data.
5. This is a best-efforts system, not a privilege shield. We provide tools that help attorneys meet their duty of reasonable care. We do not — and cannot — guarantee that use of this system will prevent a finding of privilege waiver in any particular proceeding. Courts evaluate privilege claims based on the totality of circumstances, including the attorney's independent judgment about what safeguards were appropriate for the matter at hand.

11.4 Your Responsibilities

You acknowledge and agree that:

1. Professional judgment. The decision to upload client materials to this Service, and to use AI-assisted analysis of those materials, is an exercise of your independent professional judgment. You are solely responsible for determining whether such use is appropriate given the nature and sensitivity of the matter, the client's expectations, applicable rules of professional conduct, and any court orders or protective orders that may restrict the disclosure of case materials.
2. Client consent. Where applicable rules of professional conduct or ethics opinions require client consent for the use of AI-assisted tools in connection with a representation, you are solely responsible for obtaining such consent before uploading client materials. We do not verify whether you have obtained client consent.
3. Supervision of output. All AI-generated analysis, objections, discovery requests, and other work product require your independent review and professional judgment before use. You are responsible for verifying the accuracy, completeness, and appropriateness of all output.

11.5 Limitation of Liability

To the fullest extent permitted by applicable law, the operators of this Service shall not be liable for any claim, loss, damage, disciplinary proceeding, malpractice action, bar complaint, or other consequence arising from or relating to: (i) an alleged breach of attorney-client privilege, work product protection, or duty of confidentiality resulting from your use of the Service; (ii) the failure of our entity obfuscation system to detect or replace any particular item of identifying information; (iii) any determination by a court, tribunal, or disciplinary body that your use of AI-assisted tools was inconsistent with your professional obligations; or (iv) the retention, processing, or potential exposure of case data by our third-party AI provider during the data retention period described in our Privacy Policy.

11.6 Indemnification

You agree to indemnify, defend, and hold harmless the operators of this Service from and against any claims, damages, losses, costs, and expenses (including reasonable attorneys' fees) arising from or relating to your use of the Service in connection with a client representation, including any claim that such use breached a duty of confidentiality, privilege, or professional conduct.

11.7 Acknowledgment

By using the LITIGAGENT case workspace tools, you acknowledge that you have read and understood our Privacy Policy (including the description of entity obfuscation, data flows, and third-party processing), that you understand the limitations of automated obfuscation as described herein, and that you accept full responsibility for determining whether the safeguards we provide are sufficient for your particular use case.