Privacy policy
What Passim collects, why, and how long it keeps it.
Last updated: November 9, 2026
Passim is a legal research service operated by Passim AI LLC. This policy explains what information Passim collects when you use the website, the API, and the Passim MCP server, why it collects it, and how long it keeps it.
Two things frame everything below. Passim is a retrieval service: the AI assistant you use sends it a research query and receives legal sources back. Passim receives what is included in those requests. It does not automatically receive the rest of your conversation or your files, and your assistant decides what goes in a query, so keep client identifiers out. And Passim does not use anything you send it to train or fine-tune any model.
What we collect
| Information | Why we have it | How long we keep it |
|---|---|---|
| Account details | Your email address, name, sign-in identity from your OAuth provider, and sign-in history, so we can give you access and secure it. | While your account is open, then deleted on request. |
| Access credentials | API keys (stored hashed), OAuth claims, and the tools each key may call. | Until you or we revoke them. |
| Research activity | Which tools were called, by whom, when, and how often, so we can run the service, enforce rate limits, and bill per seat. The text of your queries is not retained. | Operational logs 90 days; usage counts while your account is open. |
| Email alert settings | The practice areas, circuits, sources, and schedule you choose for email alerts, the cases, policies, and lawsuits you follow, and a record of the alerts we sent you, so each alert matches your choices and does not repeat. If an alert bounces or is marked as spam, our email provider tells us and we stop sending to that address. | While your account is open, then deleted on request. |
| Messages to us | Access requests, support, privacy, and security messages, including your name, firm, and email. | Kept as business records while relevant. |
| Device and usage data | IP address, browser or client type, pages and routes requested, timing, and errors, for security and reliability. | 90 days. |
| Public website analytics | Page views, referring website domains, browser and device categories, selected campaign tags, and clicks on connection, trial, plan-selection, and billing links, to understand visits to our home, library, pricing, and Get started pages. | Up to one year. |
| Billing details | Plan, seats, and payment status. Stripe handles payments, subscriptions, and tax. Card numbers go to Stripe and never reach Passim’s servers. | As required for tax and accounting records. |
Please do not put Social Security numbers, A-numbers, passport numbers, dates of birth, or other client-identifying facts into a research query. Passim does not need them to find the law.
Your queries
The content of a query, meaning your search terms, citations, and questions, is processed to answer it and then discarded. Passim’s usage records keep the tool name, the calling account or key, counts, timing, and dates. Queries travel in request bodies rather than URLs, so they never appear in access logs, and our error reporting is configured not to capture request bodies. When an operational problem needs correlation, content fields are fingerprinted rather than stored.
If your firm’s plan covers your seat, the person who administers that plan can see your roster email address. If you first joined Passim through that firm plan, the administrator can also see whether you have signed in and the date you last used Passim. Joining an existing personal account to a firm plan does not share that account’s sign-in or usage activity with the administrator. Administrators never see your queries.
Your AI assistant is a separate service
Claude, ChatGPT, or whichever MCP client you use processes your conversation, decides when to call Passim, and receives Passim’s results. It may keep that material under its own terms and your workspace’s settings. Passim does not control that provider’s retention, training, human review, data residency, or administrator settings, and its promises here apply only to Passim. For confidential client work, use a workspace whose agreement meets your firm’s requirements.
No model training
Passim does not use account data, queries, tool calls, or messages to train or fine-tune any model, and does not send your queries to any AI provider. Where Passim uses AI itself, it is to summarize and classify public legal materials, never your data.
Cookies and analytics
Render’s hosting for this site, the account pages, the reader, and the connector runs behind Cloudflare, which terminates TLS for that traffic. Our contact and newsletter forms use Cloudflare Turnstile to prevent automated spam. Cloudflare processes browser and device signals for verification. We send the verification token to Cloudflare, not the name, email address, firm, or message you enter. See Cloudflare’s privacy policy.
Passim uses cookies for sign-in, session security, and rate limiting, one that remembers your setup choices during sign-up for up to eight hours, and one that remembers a ref label on the account pages. If a sign-up link carries a label we recognize, the account pages keep it in your browser for up to seven days and add one to a daily total for that label when you arrive, create an account, or complete checkout. The totals aren’t connected to your account, email, or IP address, and the label isn’t stored with your account or sent to Stripe. We use PostHog for cookieless visitor analytics only on the public home, library, pricing, and Get started pages at passimai.com. There is no advertising, no cross-site tracking, and no marketing analytics on any research, account, reader, or MCP surface, apart from those daily totals.
The PostHog analytics count page views and selected connection, trial, plan-selection, and billing-link clicks, together with referring website domains, browser and device categories, and a limited set of campaign tags. Approved campaign tags can travel in links between public pages so a visit retains its campaign label; this does not store campaign tags in cookies or browser storage, and we do not add them to account, research, or external links. These click counts measure interest, not completed registration or payment. We exclude raw URL query strings and fragments, referrer paths, form entries, research queries, client information, and account identifiers. We do not create visitor profiles or record sessions, screens, keystrokes, or page content.
PostHog receives the IP address and browser user-agent needed for the network request and uses them, with our website hostname and a daily changing salt, to estimate daily visitors. IP addresses are removed from stored analytics events. The analytics do not store visitor identifiers in cookies, local storage, or session storage, and do not link a visitor across days or devices. We skip analytics when the browser sends Global Privacy Control or Do Not Track. PostHog processes these limited analytics in the United States, with event retention of up to one year.
On passimai.com, if you arrive through a link carrying a short ref label, such as the name of a talk or post, the page keeps that label in your browser’s session storage until you close the tab and adds it to our sign-up links. The label describes the link, not you, and PostHog does not receive it.
How information is shared
Passim shares information with a small number of providers that help run the service, such as hosting, sign-in, payments, email delivery, and error monitoring. Each is listed on the subprocessors page with what it does and where it is, and each processes information only as needed to provide its service. Passim does not sell personal information and does not share it for advertising.
Passim may also disclose information when the law requires it, to protect the security of the service or the rights of others, to investigate misuse, or as part of a merger, acquisition, or financing, in which case this policy will continue to apply to the transferred information.
Public legal records
The Passim library is built from public legal materials. Those records can contain personal information because a court or agency published it. Passim uses them for research, citation, and source verification. If a record returned by Passim exposes information you believe should be restricted, write to us with the source URL, the document identifier, and the concern, and we will review it.
Retention and deletion
Operational logs are deleted after 90 days unless a specific security incident, abuse investigation, dispute, or legal obligation requires keeping particular records longer. Account data is kept while your account is open and deleted when you ask us to close it. Business records and public legal materials are kept as long as they are needed for the purposes described above.
Security
The controls that protect your information are described on the security page, including what is in place, what is planned, and what has not yet been independently audited. No online service can promise perfect security, and Passim does not.
International users
Passim is operated from the United States. Our providers may process information in the United States and other countries, including where Cloudflare operates its network and verification service.
Contact
Passim AI LLC. Privacy questions and requests: chris@passimai.com.