Legal

• Identity and Contact Data: We collect your name and email address when you create a CATO account. We may also collect or generate indirect identifiers (e.g., internal user IDs).
• Payment Information: If you purchase access to paid features, we collect payment information through our payment processor. Nyrus does not store full payment credentials on its servers.
• Inputs and Outputs: Your messages and the responses and analysis results they generate. Input, Output, and User Content have the meanings given in Section 6 of the Terms of Service. If you include personal data in your Input, it will be processed as part of the conversation and may appear in Output.
• Uploaded Datasets: Files you upload for analysis, including their contents, filenames, and structural metadata (column names, row counts, data types).
• Literature Search Data: Search queries you submit through Literature Intelligence, which are transmitted to external academic databases. This category also includes literature collections you create, annotations you add to papers, saved watch alert queries, and papers you send to Chat via the literature-to-chat bridge.
• Memories: A small, user-controlled index of durable context, such as preferences, standing instructions, project definitions, constraints, and decisions. Automatically created memories must be grounded in a user message. Generated scientific findings are not automatically promoted to memory. Each memory records its scope and source so you can review, edit, pin, or delete it. When automatic memory is enabled, CATO reviews changed conversations in a daily batch. The batch contains human-readable conversation excerpts and existing memory text, but excludes uploaded files, figures, binary artifacts, recognized credentials, and sandbox storage access.
• Feedback and Communications: If you submit feedback or communicate with us, we collect the contents of those messages.
• Service and Agent Trajectories: We record prompts, responses, tool calls and results, generated code, citations, reviewer results, execution metadata, safety signals, provider-returned reasoning content where available, and User Content surfaced in those records. Uploading a file does not, by itself, place the full raw file in a trajectory. File excerpts, selected rows, summaries, tool results, generated outputs, and other content surfaced to the agent may be recorded.

When you use the Services, we receive certain technical data automatically, including device type, browser information, IP address, and usage information such as dates and times of access.

CATO uses a small number of cookies and browser storage entries, all of them to run the Services rather than to track you. A session cookie keeps you signed in. If you choose to trust a device during two-factor authentication, a cookie remembers that choice; this option is not available to members of an enterprise organization. Nyrus administrators may carry a cookie that lets them test the Services as a different plan. Browser storage holds interface preferences such as column widths, panel state, and the tab you last used. We do not use advertising cookies, and we do not currently run third-party analytics on the Services. If that changes, we will update this Policy and, where the law requires it, ask for your consent first. You can clear cookies and browser storage in your browser settings; doing so signs you out and resets your preferences.

If you join a lab, the conversations, files, and outputs in projects assigned to that lab are visible to the other members of the lab and to its owner and admins. If you create a share link for a conversation or figure, anyone with the link can view that content until you revoke the link. Projects you keep outside a lab, and content you do not share, are visible only to you and to Nyrus for the purposes described in this Policy.

A lab's owner and admins can see the conversations, files, and outputs in projects assigned to the lab, manage its membership and seats, and delete lab projects. They cannot see projects you keep outside the lab. Where your organization has an enterprise agreement with Nyrus, it may also receive account, usage, and billing information about its members and may ask us to suspend or remove a member's access. If you joined through your organization, its administrators may be told that you hold an account. Nyrus personnel access your data only to operate the Services, resolve a support request you raise, investigate abuse or a security incident, or as required by law.

We may disclose personal data to a buyer or successor in connection with a merger, acquisition, financing, reorganization, bankruptcy, or sale of all or part of our business or assets, including during due diligence for such a transaction. The successor will be bound by this Policy for the data it receives, or will tell you otherwise.

We may also disclose personal data to courts, law enforcement, regulators, or other third parties when we believe in good faith that the law or legal process requires it. We may do the same to enforce our Terms of Service or Usage Policy, to detect or prevent fraud, abuse, or a security threat, or to protect the rights, property, or safety of Nyrus, our users, or the public. Where the law allows, we will tell you before disclosing your data in response to a legal request.

Uploaded files, project artifacts, and each project's durable working filesystem are stored with our cloud storage provider. Account, conversation, project, memory, and literature records live in CATO's application database, hosted with our application hosting provider alongside caches, workflow state, and service logs. Encrypted backups and the error-monitoring and payment systems described below also hold copies of some data.

When the agent works on a project, its filesystem is attached to an isolated sandbox. The sandbox receives no cloud-storage credentials or application secrets, cannot reach other projects' filesystems, and is blocked from internal cloud networks and metadata services. Files stay in the project filesystem until you delete them or the project. Sandboxes may reach an approved list of public destinations over HTTPS, such as package repositories and scientific databases, when an analysis, package installation, or tool needs it; data your code sends to one of them is subject to that destination's terms.

Trajectories are recorded as described in Section 1. Uploading a file does not copy its full contents into a trajectory, although excerpts, tool results, and outputs surfaced to the agent may be recorded. Operational trajectories and training-approved trajectories are kept in separate stores: production systems can create operational records but cannot read them back, and training systems can read only records that a separate cleaning and approval step has copied into the training-approved store. Accounts on plans that are exempt from model-improvement capture do not produce full-content trajectories at all.

You can delete conversations, memories, datasets, files, and projects in the application, and you can delete your account from Settings. Deletion removes the primary application copy; security, diagnostic, trajectory, and backup copies may persist for the periods below.

• Account and billing records: for the life of the account, then removed on account deletion except where payment-processor or tax rules require longer retention.
• Conversations, memories, datasets, project files, outputs, literature collections, annotations, and watch queries: until you delete them or delete your account.
• Literature full text: retrieved on demand; a paper is stored with the project only when you or CATO saves it to the workspace, and stays until you delete it or the project.
• Agent trajectories: an event-level record of each turn's tool calls is stored with the conversation and deleted with it. Full-content trajectories, where created, are retained as reasonably necessary to operate and restore workflows, debug and evaluate CATO, prevent misuse, investigate security events, resolve disputes, enforce our policies, and improve our offerings.
• Error-monitoring events: for the retention period configured with our error-monitoring provider, then discarded.
• Technical logs and backups: on a rolling basis for operational recovery, after which they are overwritten or expire.

When you delete content or your account, the primary copy is removed at once, and copies held in caches, diagnostic systems, and security systems are removed within 30 days. Backups expire on their rolling schedule, and we do not restore deleted content from a backup except to recover the Services after an outage. Deleting your account removes your projects, conversations, memories, files, and literature records on the same timeline. We keep a record of the deletion request itself so that we can show it was honored.

Records connected to abuse, security, fraud, disputes, or legal obligations may be retained longer when reasonably necessary for those purposes. Operational trajectory records are written to a store that cannot be edited or deleted record by record, so a trajectory that has already been recorded is not removed on request. You can email contact@nyrus.ai to ask that identifiable records no longer be selected for future model-improvement datasets. A request applies once processed and does not undo completed processing; de-identified or aggregated data may be retained where the law permits, and changes already incorporated into trained model weights cannot generally be reversed.

CATO can save durable context you state in a conversation and recall a relevant subset later. Global memories can apply across projects; project memories stay within their project. Automatic saving, recall, and cross-project recall each have their own control, and every memory can be viewed, edited, pinned, or permanently deleted from Settings > Memory. Memory is not treated as scientific evidence, and current instructions override older memories.

The hosted Services rely on contracted providers for AI inference and memory maintenance, application hosting and databases, file storage and sandbox infrastructure, error monitoring, payment processing, and email delivery. We limit each provider to the data its function needs, and we do not authorize AI inference or retrieval providers to train their models on your User Content. The function, data categories, retention posture, and contract for each provider are kept current on our Subprocessors and Service Providers page, which we update before a new provider receives production user data.

Literature, clinical trial, patent, protein, pathway, chemical, and other public research databases are external data sources rather than contracted providers. CATO sends them only the search terms, identifiers, and filters a lookup needs, and their own terms and logging practices apply. AI models and infrastructure are upgraded periodically, so your data may be processed by updated model versions.

We protect your data with encryption in transit and at rest, sandboxed code execution that receives no storage credentials and cannot reach internal cloud networks, separate identities for user files and for each trajectory store, signed access to the API, and two-factor authentication, which you can enable on your account and which is required for members of an organization with an enterprise agreement. Section 7 of the Data Processing notice lists the measures in place. No security system is impenetrable, and we cannot guarantee absolute security.