Privacy Policy

Empyralis Privacy Policy

Empyralis is a workspace for a team’s projects, tasks, and documents, alongside AI agents that can act on that team’s behalf — including, when configured, running commands on a computer the team owns or provisions. This policy explains what information we collect to operate Empyralis, why we collect it, and who else sees it.

Last updated: August 27, 2026.

This document was drafted with AI assistance and has not yet been reviewed by a lawyer. It is meant to be an honest, accurate description of what Empyralis actually does with your data today — not a legally sufficient privacy notice. Do not rely on it as final until it has been reviewed by qualified counsel, and treat every bracketed note below as an open item.

Who operates Empyralis

Empyralis is operated by its founder, who is based in Uzbekistan.

[The specific legal entity that will operate Empyralis, its registered address, and its jurisdiction of incorporation have not been established yet and are not listed here. This section needs to be completed once that entity exists, and reviewed by a lawyer for what it implies about which country’s privacy law applies to you.]

Information you provide

When you create an account, we collect your name, email address, and password (stored as a salted hash, never in plain text) — or, if you sign in with Google, the identity information Google provides. When you use the product, we store the projects, tasks, documents, and comments you and your teammates create, and any files you deliberately upload to a project (notes and images only — the upload system rejects executable code, archives, and binary files, checking the actual file content rather than trusting its name or extension).

What an agent generates and remembers

When you configure an agent and it works on a task, we store the messages exchanged through the channel you connected, the agent’s replies, records of the tools and commands it ran and their outcomes, and — where you or the agent choose to save one — a memory note about the project or a private note tied to your own account. These records live on our own servers; today that is a single server located in San Francisco, California, alongside a managed database that stores your workspace’s structured data (accounts, tasks, documents, and similar records).

Agents running commands on your computer

If you attach an agent to a computer you own or provision, that agent can execute shell commands and read or write files on that machine, at your instruction or the instruction of someone you allow to message it. We do not store a general history of your computer’s files; a summary of the specific commands run and their results is recorded as part of the agent’s activity log described above, so that you can review what happened. Attaching hardware is a decision only the hardware’s owner makes, and it is never inherited automatically by other members of a workspace.

Credentials and connected accounts

When you connect a third-party account — an OAuth connection to an app like Google or Notion, an API key for an AI model provider, or a bot token for a messaging channel — we store what is needed to maintain that connection separately from ordinary workspace content, encrypted at rest, and use it only for the actions you configure or explicitly request.

One category of credential is different: for channels that run through a computer you have paired (for example, WhatsApp, Signal, or iMessage), the credential you paste in is sent directly to that computer and applied there. It passes through our servers but is not stored, logged, or retained by us at all — only its owner’s own device holds it.

Third-party AI model providers

When an agent responds to you, the content of your conversation and relevant workspace context is sent to an AI model provider to generate that response. Depending on how the agent is configured, this may be a provider we operate on your behalf (for example, DeepSeek), or a provider you have connected with your own account, subscription, or API key (which today may include Anthropic’s Claude, OpenAI, xAI’s Grok, Google’s Gemini, and others, or a coding-assistant subscription such as Codex or Cursor). That provider processes the content under its own privacy practices, which we do not control and which you should review directly. We do not sell this data and do not use it to train our own models.

Messaging channels

If you connect a messaging channel — such as Telegram or Slack, or, for agents attached to a paired computer, WhatsApp, Signal, iMessage, or a similar platform — messages sent to and from your agent pass through that channel’s own service and are subject to that provider’s own privacy practices in addition to this policy. We store the message content and delivery records needed to run the conversation and show it back to you in Empyralis.

Google user data

Empyralis uses Google user data only after you grant consent through Google OAuth, and only to carry out the connected-app actions you authorize — for example, reading, drafting, or sending Gmail, or reading or creating Google Calendar events, when you ask an agent to do so. Empyralis does not sell Google user data, does not use it for advertising, and does not use it to train AI models.

Billing information

Credit purchases are processed by Polar, acting as our merchant of record. We receive confirmation that a payment was made and the amount, but Polar — not Empyralis — collects and stores your payment card details.

Why we use this information

We use the information above to operate the Service: authenticating you, running agents and delivering their replies, enforcing usage limits and storage caps, billing for credits, keeping the transparent record of tool calls and agent activity that Empyralis is built around, responding to support requests, and protecting the platform from abuse.

International data transfer

Empyralis is operated from Uzbekistan, our servers are located in the United States, and some third-party providers we use (including our payment processor and some AI model providers) are also U.S.-based. Using Empyralis means your information may be processed in these locations.

Retention and deletion

You can delete individual tasks, documents, and connected accounts from within Empyralis. Disconnecting an account (for example, a Google connection) stops us from making new API requests through that connection. For a personal messaging channel, you can request deletion of your conversation history by sending /delete to the agent in that channel.

[Specific retention periods for each category of data — how long a deleted workspace’s records are kept before permanent removal, how long billing and audit records are retained for legal reasons, and how a full account-deletion request is handled — are not finalized yet and should be defined before this policy is relied upon.]

Children’s privacy

Empyralis is not directed to children and is not intended for use by anyone under 18.

Your rights

Depending on where you live, applicable law may give you rights to access, correct, delete, or export your personal information, or to object to certain uses of it. You can exercise many of these directly in the product; for anything else, contact us using the details below.

[Whether GDPR, CCPA, or another specific privacy law applies to you, and the exact process and timeline for exercising rights under it, has not been determined and should be added once Empyralis’s legal entity and customer base are established.]

Changes to this policy

We may update this policy as the product changes. If we make a material change, we will update the date above and, where practical, let you know.

Contact

Questions or privacy requests can be sent to mansurao886@gmail.com.