How VeriLex handles your data — and how it doesn't.
Last updated: June 4, 2026
VeriLex (verilex.org) is a free legal document accessibility tool built and maintained by Tera Chen, a Legal Studies student at the University of California, Berkeley. VeriLex is not a law firm, legal services company, or registered legal entity.
Contact: terachen@berkeley.edu
Text you paste or files you upload are transmitted securely (HTTPS) to Anthropic's Claude API for analysis. VeriLex never stores, logs, or retains document content on its own servers. The analysis result is displayed in your browser and discarded when you close or navigate away from the page.
VeriLex logs a small set of anonymized metadata for each analysis to monitor service health and improve the tool. Each log entry contains:
This log contains no document content, no IP addresses, no names, and no personal identifiers. It is held in server memory and is cleared automatically each time the server restarts (which happens with every new deployment).
If you submit feedback through the feedback form, the message, category, and any name or email address you optionally provide are stored in server memory. This data is also cleared on server restart. We do not share it with any third party.
VeriLex uses no cookies, no tracking pixels, no analytics scripts, and no third-party tracking technology of any kind. Your preferred interface language is saved in your browser's localStorage — this data never leaves your device.
Your document text is processed by Anthropic's Claude API. Anthropic is a US-based AI safety company. Per Anthropic's privacy policy and their API usage terms, Anthropic does not use API inputs to train their AI models by default. For Anthropic's full data practices, see anthropic.com/legal/privacy.
VeriLex is hosted on Railway, a cloud infrastructure provider. Servers are located in the United States. For Railway's privacy practices, see railway.app/legal/privacy.
VeriLex offers an optional Google Drive integration. If you choose to use it, you will be asked to sign in with your Google account and grant VeriLex read-only access to files you select. This access is used solely to download the specific file you choose so it can be analyzed.
VeriLex does not store your Google account information (name, email address, or profile data), does not access any files you did not explicitly select, and does not retain access to your Google Drive after the session ends. The OAuth access token issued by Google is held in browser memory only and is discarded when you leave the page.
The content of the file you select is processed exactly like any directly uploaded file — sent to Anthropic's Claude API for analysis and not stored on VeriLex's servers. For Google's data practices, see policies.google.com/privacy.
Use of Google Drive through VeriLex is governed by the Google API Services User Data Policy, including the Limited Use requirements.
VeriLex is designed to help users understand documents that may be sensitive — including immigration notices, eviction filings, and financial correspondence. If your document contains highly sensitive personal information (Social Security numbers, immigration case numbers, full legal names combined with dates of birth), we recommend redacting that identifying information before uploading. The legal substance of the document is usually analyzable without it.
While VeriLex does not store documents, any transmission to an external AI service carries inherent risk. Use your judgment about what you share.
VeriLex is not directed at children under 13 and does not knowingly collect information from children under 13. If you believe a child under 13 has submitted information, please contact us at terachen@berkeley.edu.
This policy may be updated over time. The "Last updated" date at the top reflects when it was last revised. Continued use of VeriLex after changes are posted constitutes acceptance of the updated policy.
For privacy questions, concerns, or requests: terachen@berkeley.edu