Privacy policy

The short version

• This website sets no cookies, runs no analytics, and loads no outside scripts. If you contact us, we use what you send to reply to you.
• In the product, Mooseworks reads the email and documents of a transaction from mailboxes and accounts your organization connects, so it can answer routine requests, keep status current, and flag what needs a person.
• We are not the owner of that material. Your brokerage or team is. We handle it on their instructions and for no other purpose. We do not sell personal information, and we do not use your data to train AI models.
• Nothing is sent on your behalf without a person approving it, and every action the system takes is written to an audit log you can read.

Who we are

Mooseworks operates this website (https://mooseworks.ai) and the Mooseworks application at https://app.mooseworks.ai.

Two different relationships

This policy covers two situations that are worth keeping apart, because your rights differ between them.

You visit this website or contact us. We decide what happens to that information, and it is described under “This website” below.

Your organization uses the product. The email, documents, and contact details inside a workspace belong to the organization that connected them. We act as a service provider under the applicable US state privacy laws: we process that material on the organization’s instructions and do not use it for our own purposes. Where the GDPR applies to a customer, we act as their processor on the same terms. If you are a buyer, seller, or other contact to a transaction and want to know what a brokerage holds about you, that request goes to the brokerage; we will help them answer it.

This website

Every page here is static HTML generated at build time. There are no cookies, no analytics or advertising tags, no embedded fonts or scripts from other companies, and no tracking pixels. Nothing about your visit is shared with anyone.

Our web server keeps ordinary access logs — IP address, date and time, the page requested, the referring page, and your browser’s user-agent string. We use them to keep the site up and to investigate abuse, and for nothing else. They are kept for a short period and then discarded.

If you request early access or otherwise get in touch, we receive whatever you send: your name, email address, and anything you choose to tell us about how your transactions run. We use it to reply, and to keep track of the conversation while it is live. We will not add you to a marketing list you did not ask for.

What the product processes

When your organization connects an account, Mooseworks reads from it on a schedule and stores what it reads so it can reason across a whole transaction. In practice that means:

• Messages and documents. Email in the connected mailboxes — headers, bodies, and attachments — plus files imported directly. Each stored item records which connection it arrived through.
• People. Names, email addresses, phone numbers, and roles of the contacts to a transaction, as they appear in that correspondence: clients, agents, lenders, title, and inspectors among them.
• What the system concluded. Extracted dates, statuses, and flags, each pointing back to the message or document it came from and the model version that produced it. These are never edited in place: a new conclusion supersedes the old one, and the history stays readable.
• Account and usage records. Who signed in, from which identity provider, and an audit event for every state change the system makes.

Correspondence about a home purchase can contain sensitive material — financial details, identity documents, and sometimes health or family circumstances behind a move. We do not seek it out, we apply the same protections to all of it, and we do not use any of it for profiling, advertising, or any purpose other than coordinating the transaction for the organization whose workspace it is in.

AI models

Reading a message and extracting what it means is done by a large language model. Message and document text is sent to Anthropic’s API for that purpose. Under Anthropic’s commercial terms, material sent through the API is not used to train their models. We do not train models on your data either, and we do not allow anyone else to.

The model does not act on its own. It reads and it proposes; a person approves anything that leaves the system, and the model is never the one to compute a deadline — it records what a document says and the dates are derived from that. Every model call is logged with its inputs and the version that ran, so a conclusion can always be traced back.

Who else sees data

We do not sell personal information, and we do not share it for cross-context behavioural advertising. We use a small number of vendors to run the service:

• Amazon Web Services — hosting, databases, and file storage, in the US East (Ohio) region.
• Anthropic — model inference, as described above.
• Microsoft and Google — only where your organization connects their accounts: to sign you in, and to read the mailboxes you point us at.
• Dotloop — only where your organization connects a dotloop account: to read loop, document, and contact data for active transactions.

Each is bound by terms that limit them to providing that service. We will also disclose information if the law requires it, and we would tell the affected customer unless we were forbidden to. If the business is ever sold, data moves with it and this policy travels along until it is replaced by one no less protective.

Where data is processed

Data is processed in the United States. If we serve customers or data subjects in the UK or the EEA in future, we will update this policy with the safeguards that apply to those transfers.

How it is protected

• No passwords. Sign-in is single sign-on through your Microsoft or Google work account. We never see or store a password.
• Credentials are encrypted before they are stored, with keys held outside the database.
• Workspaces are isolated in the database itself, not only in application code: every record carries its workspace and every query is scoped to it.
• Every change is audited in the same transaction that makes it, so the log cannot silently disagree with the data.
• Traffic is encrypted in transit with TLS.

No system is perfectly secure, and we would rather say that than imply otherwise. If you find a vulnerability, please report it to us using the contact details below.

Keeping and deleting data

How long we hold a workspace’s material depends on the state the work is in.

• Open transactions. While a transaction is open we keep its content in full — messages, documents, and everything extracted from them. The system’s value is in reasoning across a transaction’s whole history, so there is no useful way to hold less.
• Closed transactions, in an active workspace. Twelve months after a transaction closes, we purge the underlying message and document content. What remains is the extracted record — dates, statuses, and the audit trail behind them — which is structured data about the deal rather than the raw correspondence it came from.
• A terminated workspace or a disconnected account. Disconnecting an account stops further reading immediately and revokes our access to it. When a workspace is terminated we delete all of its data within 30 days, and sooner on request from the organization that owns it.
• Audit events and model-call records. These are append-only and kept for the life of the workspace. They are the accountability mechanism — what lets you check what the system believed and what it did — not raw personal content.

Your rights

Under the California Consumer Privacy Act as amended by the CPRA, and the comparable privacy laws of other US states, you may have the right to know what personal information we hold about you, to get a copy of it, to have it corrected or deleted, to limit how it is used, and to be free of discrimination for asking. Where other privacy laws apply to you, we will honour the equivalent rights they give you.

Ask using the contact details below and we will respond within the time the applicable law allows — for most requests, 45 days or sooner. Where the material sits in a customer’s workspace, we will pass the request to that organization and help them act on it, because it is theirs to answer.

Mooseworks makes no automated decision that has a legal or similarly significant effect on anyone. The system proposes; people decide.

Children

This is a tool for real estate professionals. It is not directed at children, and we do not knowingly collect information from anyone under 16. If you believe we have, tell us and we will delete it.

Changes to this policy

When this policy changes we will update the effective date at the top, and for anything that materially changes how we handle personal information we will notify the workspaces affected before it takes effect.

Contact

Mooseworks is in early access and not yet open to the public. Our legal entity details, postal address, and the address for privacy requests will be published here before this policy takes effect.