Privacy Policy

This Privacy Policy explains what information Corpus (“Corpus,” “we,” “us”) collects when you use corpuslaw.us, the Belisarius formation assistant, and our developer surfaces (the public API and the Model Context Protocol server), how we use and share it, and the choices you have. It works alongside our Terms & Disclosures. Corpus is a software platform, not a law firm, and does not provide legal advice.

Information we collect

Formation intake. To prepare and file a business entity, we collect what you provide in the intake: the entity type and state, the proposed company name, the principal office address, a NAICS industry code, and the names, roles, and addresses of the members, managers, officers, directors, organizers, and incorporators, along with contact details.

EIN and Social Security Number. Only if you add the optional EIN service, we collect the SS-4 answers and the responsible party’s Social Security Number. The SSN is entered into a dedicated secure field — it is never sent to the AI assistant or to any AI provider, never appears in the human-review snapshot (a masked last-4 is shown instead), and is permanently deleted from our systems as soon as the EIN filing is submitted, and in every case within 7 days of when you provide it — whichever comes first. The 7-day limit is enforced by an automated nightly deletion job, so the SSN is erased even if the filing stalls or is never completed. If that happens before we file, we will ask you for it again rather than keep it.

Account. If you sign in, we collect your name and email address through Google Sign-In (managed with Better Auth). An account lets us persist your orders and your conversations with the assistant.

Payments. Card payments are processed by Stripe. When you pay in USDC on Solana, Corpus receives the transfer into an entity-controlled wallet and records the recipient, mint, cluster, transaction signature, amount, and settlement state needed to reconcile it. We do not receive or store private wallet keys. The wallet address and transaction signature are public-ledger data and cannot be made private or deleted from the ledger; do not include personal information in a transaction memo. A USDC transfer is not treated as credited until the displayed recipient, mint, cluster, amount, and a finalized matching transaction are verified.

Conversations and memory. We store the messages you exchange with the Belisarius assistant (formation intake and legal search) and, for signed-in users, a short running summary that provides continuity across sessions.

Developer / API and MCP usage. If you create an API key or connect an agent to our MCP server, we collect the account and key metadata and log request details (such as the tool called and timestamps) to meter usage, enforce rate limits, and secure the service.

Automatically collected. Like most websites, we collect usage and device information (pages viewed, approximate location from IP, browser type) and use cookies and similar technologies for authentication, security, and analytics.

How we use your information

We use your information to provide the services — running the intake, preparing and submitting filings after human review, obtaining an EIN if requested, servicing registered agent appointments, and answering legal-search questions; to process payments; to communicate with you about your order and provide support; to secure, debug, and improve the platform; and to meet our legal, tax, and accounting obligations. We do not use your Social Security Number for anything other than obtaining the EIN you requested.

The AI assistant and AI providers

Belisarius is powered by artificial intelligence. The text you type into the assistant — and the intake details it works with (excluding the SSN, which is never sent to any AI provider) — is transmitted to third-party AI providers to generate responses and to search the law: the conversational model runs on Alibaba Cloud’s DashScope (Qwen); text embeddings for legal search are generated by Voyage AI; and the cross-session memory summary is generated through OpenRouter / NVIDIA NIM. These providers process the data to return a result to us; some may process it outside the United States. If you would rather not share particular details with an AI model, do not enter them into the chat — you can complete an order through structured fields instead.

How we share your information

We share your information only as needed to run the service, with these categories of providers (subprocessors):

• Filing partner and government. Filing and registered agent details are shared with Corporate Tools / Registered Agents Inc. and with the relevant Secretary of State (or the IRS for an EIN). Formation filings generally become public record with the State.
• Payments. Stripe, Inc., to process card payments. A USDC payment on Solana is visible on the public ledger; Corpus uses the configured Solana RPC service to read transaction metadata for reconciliation.
• Infrastructure. Amazon Web Services (Aurora PostgreSQL database, United States region) and Vercel (application hosting and Vercel Analytics).
• AI providers. Alibaba Cloud (DashScope / Qwen), Voyage AI, and OpenRouter / NVIDIA NIM, as described above.
• Communications, monitoring, and analytics. Resend (transactional email), Sentry (error monitoring), PostHog (product analytics), and Google (sign-in).

We may also disclose information if required by law or to protect our rights, users, or the public, and in connection with a merger, acquisition, or sale of assets (with notice). We do not sell your personal information, and we do not share it for cross-context behavioral advertising.

Third-party agents, the API, and the MCP server

Corpus can be reached by AI agents through our public API and our Model Context Protocol (MCP) server. When you (or an agent acting for you) call these surfaces, we process the data the agent sends to fulfill the request. The formation handoff link carries prefilled details in the URL fragment, which browsers do not send to our servers, so that personal details are not written to server logs. If you use a third-party AI client (for example, a chat app or coding agent) to interact with Corpus, that client’s own collection and use of your data is governed by its provider’s privacy policy, not this one.

Cookies and analytics

We use strictly necessary cookies for authentication and security, and privacy-bounded product analytics (PostHog and Vercel Analytics) to understand how the site is used. Our analytics events are designed to exclude personal details — we do not put names, addresses, order identifiers, free-text, or prices into analytics. Most browsers let you block or delete cookies; blocking necessary cookies may prevent you from signing in.

Data retention

We keep order and account data for as long as needed to provide the services and to meet legal, tax, and accounting obligations, then delete or de-identify it. The responsible party’s Social Security Number is an exception — it is permanently deleted when the EIN filing is submitted, and never retained longer than 7 days regardless. Published law text and citations are not personal information and are retained as part of the corpus.

Security

We protect your information with encryption in transit (HTTPS/TLS on every connection, including between our application and our database), access controls (database access uses short-lived identity-based credentials rather than a shared password, and the staff console is restricted by account role), and per-order access tokens so one customer’s order cannot be read from another’s session.

The responsible party’s Social Security Number gets specific handling: it is never sent to the AI assistant or to any AI provider, it is masked to a last-4 in the human-review snapshot, and it is hard-deleted within 7 days at the latest by an automated nightly job — whether or not the filing completed. Every filing is human-reviewed, and the approved filing data is cryptographically hash-locked at review time — a tamper-evident record of exactly what a person approved.

No method of transmission or storage is perfectly secure, but we work to protect your information and to respond promptly to any incident.

Your choices and rights

You may request access to, correction of, or deletion of your personal information, and you may opt out of non-essential communications at any time. Depending on where you live, you may have additional rights (for example, under the California Consumer Privacy Act or the GDPR), including the right to know what we hold and the right to deletion, which you may exercise without discriminatory treatment. To make a request, email support@corpuslaw.us; we may need to verify your identity. Some data must be retained to complete a filing already submitted to a government agency or to meet legal obligations.

Children

Corpus is a business service intended for adults. It is not directed to children, and we do not knowingly collect personal information from anyone under 18.

International users

Corpus is operated from the United States and stores data in the United States. Some of our AI and infrastructure providers may process data in other countries. By using Corpus, you understand that your information may be processed in the United States and elsewhere, which may have different data-protection laws than your jurisdiction.

Changes to this policy

We may update this policy; the “last updated” date above reflects the current version. Material changes will be posted here, and where appropriate we will provide additional notice.

Contact

Questions or requests about your privacy? Email support@corpuslaw.us.