Privacy Policy

The default myAI configuration makes zero outbound network calls. Memory is a local database and a folder of markdown you can walk away with (myai memory export). We are not a processor of your context because it is never transmitted to us. There is no telemetry and no silent upload.

If you create an account on our hosted service or subscribe to a paid tier, we collect only:

• Account identity — the email address and display name you provide, and a securely hashed password.
• Workspace/tenant metadata — the tenant identifier and API-key hashes needed to scope your workspace (keys are stored SHA-256 hashed, never in plaintext).
• Billing data — handled by our payment processor (Stripe). We store a customer reference, plan tier, and subscription status; we never see or store your full card number.
• Support correspondence — anything you send us directly when you ask for help.

We do not collect your agent memory, code, repositories, prompts, or task history — those stay on your machine.

The software you self-host contacts an external service only when you supply the corresponding credential in your own .env. These are your integrations under your own accounts, not data we share:

• LLM providers (Anthropic, OpenAI, Moonshot, DeepSeek) — only with your API key, for chat/routing you invoke.
• Messaging channels (Telegram, Discord) — only with your bot token, for phone/chat control you enable.
• Error tracking (Sentry) — only if you set a DSN.
• MongoDB Atlas — only if you point your database URI at Atlas for a multi-machine queue; point it local to keep everything local.
• GitHub — the normal git workflow, driven by your own credentials.

None of these receive your memory corpus. Each provider's own privacy policy governs the data you send it.

• To authenticate you and scope your workspace.
• To provision, bill, and renew your subscription.
• To send transactional messages (receipts, security notices, service changes) — not marketing unless you opt in.
• To provide support you request and to protect the service against abuse.

We do not sell personal data, and we do not use your account data to train models.

Because your context lives on your hardware, deletion is largely in your hands: stop the stack, remove the database volume, and it is gone. For hosted account data, we retain it while your account is active and delete or anonymise it within 30 days of account closure, except where we must keep billing records to meet legal and tax obligations. To request access or deletion of hosted account data, contact us (Section 8).

Account credentials are hashed, API keys are stored SHA-256 hashed and compared in constant time, and the local stack ships loopback-bound by default. Full detail is on the Security & Data Locality page.

The site sets only functional cookies — no advertising trackers, and no analytics unless you opt in via the consent banner (none are in use today; the banner records the choice so any future tool is consent-gated):

• myai_token / myai_operator — strictly necessary: your sign-in sessions.
• myai_tenant — strictly necessary: scopes the dashboard to your workspace (an identifier only, never a credential).
• myai_locale — preference: remembers your language choice.
• myai_consent — strictly necessary: records your cookie choices so we don't re-ask on every visit.

You can change your choices at any time via the “Cookie settings” link in the footer of any policy page. Enterprise buyers: the Data Processing Addendum and the subprocessor list cover the processing side.

Questions or data requests: open a private thread via the repository or email the maintainer listed in the repo profile. We will post any material change to this policy on this page and update the “Last updated” date above.

This document is a plain-language policy, not legal advice; adapt it to your jurisdiction before relying on it commercially.