Important Notices & Legal Disclaimer

Effective date: June 10, 2026 Onix AI LLC

FDA Intel is a product of Onix AI LLC ("Onix," "we," "us"). This page explains, in plain language, exactly what FDA Intel is and what it is not. It applies to every part of the service — scores, dossiers, alerts, watchlists, AI chat answers, and AI-generated complaint drafts. By using FDA Intel, you acknowledge that you have read and understood these notices. If anything here conflicts with your expectations of the product, stop and do not rely on FDA Intel until you have resolved the difference.

These notices are written for the people who use FDA Intel: licensed attorneys and their staff at plaintiff-side law firms. FDA Intel is not a consumer service and is not directed at the general public, claimants, or anyone seeking legal help.

These notices are incorporated by reference into, and are part of, the Terms of Service between you and Onix. They supplement — and do not replace — the Terms of Service, the Privacy Policy, and any Master Services Agreement that applies to you.

1. This is an investigative tool, not legal advice

FDA Intel is an information and research tool. It aggregates public data and applies AI to help you find and triage potential litigation. It does not give legal advice, and nothing it produces is legal advice.

• No attorney-client relationship. Using FDA Intel creates no attorney-client relationship with Onix. Onix is a software vendor, not your lawyer, your co-counsel, or your expert. We do not represent you, your firm, or anyone you may represent.
• No case evaluation. A score, dossier, or draft is not an evaluation of the merits, value, viability, or settlement worth of any case. FDA Intel does not tell you whether to take a case, file a case, or decline a case.
• Your judgment governs. Every decision — whether to investigate, whom to contact, what to plead, when and where to file — is yours. FDA Intel is one input. Your independent professional judgment is the only thing that decides.

2. Signals, not conclusions

FDA Intel surfaces signals: patterns in public data that may be worth a closer look. A signal is a starting point for investigation, not a finding.

• A high score means "worth investigating." It never means "meritorious case." The MDL Hunt score, the first-mover bonus, and every weight behind them are heuristics built from public data. They measure how loud the public signals are — not whether a viable claim exists, whether causation can be proven, whether anyone was actually injured, or whether a statute of limitations has run.
• A low or absent score does not mean "no case." Real claims may produce little or no public signal, especially early. The absence of a signal is not the absence of a problem.
• Treat scores as a way to prioritize your own work, not as a substitute for it.

3. Correlation, not causation

Much of the data FDA Intel surfaces describes association, timing, or reporting — not proof of cause.

• Adverse-event data (FAERS and similar) is correlation, not causation. A report that a drug was "suspected" in an adverse event does not establish that the drug caused the event. FAERS reports are unverified, may be duplicated, are submitted voluntarily, and are subject to reporting bias. A rising count of reports may reflect more prescribing, more media attention, or more reporting — not a more dangerous product.
• "Trial terminated" does not mean "drug failed." Clinical trials are stopped for many reasons — funding, enrollment, business strategy, protocol changes, or a sponsor's commercial decision — that have nothing to do with safety or efficacy. FDA Intel attempts to flag safety-related terminations, but those flags can be wrong or incomplete. Read the underlying record before drawing any conclusion.
• A recall, warning letter, safety communication, or SEC disclosure tells you something happened in the public record. It does not tell you that any particular person has a claim.

4. AI-generated content requires attorney review

Parts of FDA Intel are generated by AI (Anthropic's Claude), including chat answers, dossier narratives, and first-draft complaint shells. AI output can be incomplete, outdated, or simply wrong, and can state things that are not true with full confidence.

• Complaint drafts are starting points, not filings. An AI-generated complaint is a rough shell to be rewritten, not a pleading to be filed. It may invent or misstate parties, dates, jurisdictions, causes of action, factual allegations, and legal citations. Onix does not warrant that any draft is non-frivolous, well-grounded in fact, or warranted by existing law.
• Verify every citation, date, party, court, and factual allegation against primary sources before any use. Do not file, serve, or send anything generated by FDA Intel without independent verification.
• Your bar duties apply in full. You remain bound by your rules of professional conduct, including the duty of competence (which extends to the technology you use), the duty of candor to the tribunal, and your obligations regarding accurate factual and legal contentions (e.g., Rule 11 and its state analogues). FDA Intel does not discharge any of these duties — you do.

5. Verify everything against primary sources

FDA Intel does not generate the underlying facts. It collects and reorganizes data from third-party public sources that Onix does not control — including FDA databases, CourtListener/PACER, FDA/CPSC/NHTSA recall feeds, SEC filings, clinical-trial registries, and news.

• That data may be inaccurate, incomplete, delayed, mislabeled, deduplicated imperfectly, or out of date. Public feeds lag, change, and contain errors. FDA Intel inherits every limitation of its sources and may add its own through parsing, matching, or scoring.
• Always confirm before acting. Confirm MDL and consolidation status directly with the JPML. Confirm dockets, parties, deadlines, and filings directly on PACER. Confirm label language, recalls, and adverse-event data against the FDA's own records. The primary source controls — never FDA Intel's summary of it.
• If FDA Intel and a primary source disagree, the primary source is right and FDA Intel is wrong. Proceed accordingly.

6. No guarantee of outcomes or completeness

FDA INTEL IS PROVIDED "AS IS" AND "AS AVAILABLE," WITHOUT WARRANTIES OF ANY KIND, EXPRESS OR IMPLIED, INCLUDING ANY IMPLIED WARRANTIES OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, ACCURACY, OR NON-INFRINGEMENT, TO THE FULLEST EXTENT PERMITTED BY LAW.

• We do not guarantee that the data is complete, current, or correct.
• We do not guarantee that FDA Intel will surface every relevant drug, product, recall, filing, or opportunity — or that any signal it surfaces will lead to a viable case or a favorable result.
• We do not guarantee uptime, that the service will be uninterrupted or error-free, or that any defect will be fixed.
• No outcome is promised or implied. Nothing in FDA Intel is a prediction or assurance of any litigation, settlement, or financial result.

7. Your responsibility

You — the user and your firm — are solely responsible for how you use FDA Intel and for everything you do with its output.

• You are responsible for independently verifying all information before relying on or acting on it. Any reliance on FDA Intel without independent verification against primary sources is unreasonable and is at your sole risk.
• You are responsible for your own compliance with all applicable laws and rules of professional conduct, including rules on competence, candor, solicitation and advertising, the unauthorized practice of law, client confidentiality, and the use of AI in legal work. Onix provides software only; its provision of the service does not constitute the practice of law.
• You are responsible for supervising your staff's use of the tool and for any work product derived from it.
• You assume all risk arising from your use of and reliance on FDA Intel. To the maximum extent permitted by law, you release Onix from, and waive, any claim arising from your use of or reliance on FDA Intel, as further limited in the Terms of Service and any Master Services Agreement.

8. How to read this page

These notices supplement — and do not replace — the Terms of Service, Privacy Policy, and any Master Services Agreement between you and Onix, which govern your use of FDA Intel, including the limitation of liability, indemnification, and the choice of Florida law and venue in Pinellas County, Florida. If you have not agreed to the Terms of Service, you are not authorized to use the service.

Onix's total liability and your exclusive remedies are limited as set out in the Terms of Service and any Master Services Agreement, including a cap on damages and the exclusion of consequential damages.

Questions about these notices: p@s.cr — Onix AI LLC, 7901 4th Street North, STE 300, St. Petersburg, FL 33702.

This page is a plain-language disclaimer provided for your information. It is not legal advice from Onix to you, and Onix is not your lawyer.