Privacy Policy

SimilarCase ("we") provides factor-based workers' compensation case-law research. This policy describes what we collect, why we collect it, and how to contact us. It is not legal advice.

What we collect

• Account details you provide (name, email, firm or role fields) when you register or sign in.
• Matter materials and case descriptions you upload or type into the product, used to extract research factors and run similar-case search for your workspace.
• Technical logs needed to operate the service (authentication events, error diagnostics, approximate usage metrics).
• Product analytics (PostHog), only after you accept the banner — page and feature events, not your confidential case text. See the section below.

Cookies, analytics, and session replay

We ask before we measure anything. Until you accept the analytics banner, no analytics script loads and no analytics cookie is set; your answer is kept in your browser's local storage, not in a cookie. You can change it at any time with the Analytics control in the site footer. Following a link from the SimilarCase app in ChatGPT sets no cookie either (see ChatGPT appbelow).

• Strictly necessary. Sign-in cookies that keep you authenticated. These are required for the workspace to function and are not covered by the banner.
• Analytics, only with consent. PostHog, hosted in the European Union, records page views, feature events (a search completed, a matter created), and errors. Person profiles are created only for signed-in accounts; anonymous visitors are not profiled.
• Session replay, only with consent. When enabled, PostHog can record a replay of the interface. Every input field is masked before the recording leaves your browser, so text you type into a matter — client facts, uploaded material, notes — is never captured. Replays are used to find broken screens, nothing else.

We do not use advertising cookies, we do not run third-party ad trackers, and we do not sell or share personal information for cross-context behavioural advertising as those terms are used in California law.

ChatGPT app

SimilarCase offers an app inside ChatGPT that finds released California WCAB decisions on a type of disputed issue. It has no sign-in and creates no account.

• What the app receives. The arguments of its one search: an issue type picked from a fixed list and, optionally, up to 80 characters of general legal terms. With each call ChatGPT also sends technical hints: anonymous identifiers for the ChatGPT account and conversation, the language, the device type and an approximate location (country, region, city).
• What it does with them. It searches released public decisions and shows up to five. Dates, docket and claim numbers, amounts and contact details are removed from the terms before the search. Please do not type names or medical details into it.
• What it does not keep. The search terms and anything else you write in ChatGPT are not stored, not sent to analytics and not used to train models. A recent result may be held in server memory for a few hours so the same search is not run twice.
• Analytics. Each search, each display of the app card and each link opened from it is recorded in PostHog, hosted in the European Union: the issue type, whether terms were given, the number of results, language, device type and the approximate location ChatGPT reports, with the ChatGPT identifiers replaced by salted one-way hashes. No search terms, no IP address and no person profile.
• Links to the site. When you follow a link from the app to similarcase.com, the site counts the visit: which app and button you came from, without your IP address or browser and without setting a cookie.
• Questions. Write to support@similarcase.com.

Your choices and rights

• Turn analytics off, or back on, from the footer control on any page.
• Ask us what account data we hold, ask for a copy, ask us to correct it, or ask us to delete it, by writing to the address below.
• California residents may exercise the rights to know, delete, and correct, and may not be treated differently for doing so.
• If you are in the European Economic Area or the United Kingdom, you may also object to processing, ask us to restrict it, and complain to your supervisory authority.

Where your data is processed

The application and its database run in the United States; product analytics runs in the European Union. Writing to us from anywhere is the same address either way.

How we use data

• To authenticate you and operate your research workspace.
• To extract factors, run corpus search, and show grounded source passages inside the product.
• To send transactional email (verification, password reset).
• To keep the service secure and debug failures.

What we do not do

• We do not sell injured workers' personal data for marketing.
• We do not use confidential matter materials for advertising or to train public models as a substitute for your workspace data.
• Insights digests and similar outreach, if offered, are opt-in.

Corpus and public sources

The legal decision corpus is built from public government decision text and related public sources under our source-governance rules. Public decision text is not your private claim file.

Sharing

We use infrastructure processors (hosting, database, auth, email, analytics) solely to run SimilarCase. Shared matter links (/s/…) expose a read-only snapshot to holders of the token; treat those links as confidential.

Retention and deletion

Account and matter data remain while your account is active. Contact us to request deletion or export of your account data. We may retain limited records required for security, abuse prevention, or legal obligation.

Contact

Privacy requests: privacy@similarcase.com

Support and the ChatGPT app: support@similarcase.com or the support page.