> ## Documentation Index
> Fetch the complete documentation index at: https://hmm.heyhyper.ai/llms.txt
> Use this file to discover all available pages before exploring further.
# Privacy & Security FAQ — How Hyper Protects Your Data
> Learn how Hyper encrypts your data, handles OAuth permissions, manages access control, and lets you delete your information at any time.
Hyper sits close to sensitive data — your emails, documents, calendars, and the institutional knowledge your team produces every day. We take that responsibility seriously. Every design decision in Hyper's security architecture starts from the assumption that your data belongs to you, and that we should handle it with the same care you would. The sections below explain exactly what we store, how we protect it, and how you stay in control.
Yes. Hyper does not sell your data to third parties — ever. Your facts, memories, and connected data are used solely to power your workspace's AI memory layer. All connector OAuth tokens are encrypted at rest using **AES-256-GCM**, which is the same encryption standard used by leading cloud providers and financial institutions. Hyper-issued bearer tokens are hashed before storage, meaning no plaintext credential is ever written to disk. Service secrets are held in **AWS SSM Parameter Store** with strict IAM access controls, so they are never embedded in application code or environment files.
Access control in Hyper is workspace-scoped. In a team workspace, each member can only read and write memories within the workspaces they belong to — there is no cross-workspace data leakage. Hyper-issued bearer tokens authenticate individual agents and users, and each token carries only the permissions granted at issuance. You can revoke any token or disconnect any integration at any time from **Settings → Connections**, and the change takes effect immediately. Administrators in a team workspace can manage member access and review connection status from the same settings panel.
Hyper stores two categories of data:
1. **Memory facts** — the structured facts your agents record via `remember` or `observe`. These persist until you delete them or disconnect your workspace.
2. **Connector data** — metadata and content ingested from your connected integrations (Gmail, Google Drive, Google Calendar, Slack). This data flows through Hyper's ingestion pipeline and is indexed to make your memories searchable.
Hyper retains your data for as long as your workspace is active. If you close your account, your data is deleted from Hyper's systems. You can also delete specific memories or disconnect individual integrations at any time without closing your account — partial deletion is fully supported.
You have several options depending on what you want to remove:
* **Delete a specific memory** — find the fact in the Hyper desktop app and delete it directly. The change is immediate.
* **Disconnect an integration** — go to **Settings → Connections**, select the integration, and click Disconnect. Hyper will stop ingesting new data from that source. You can also request deletion of previously ingested data from that connector via the same screen.
* **Delete your entire workspace** — contact Hyper support to initiate a full account and data deletion. All stored facts, credentials, and ingested data will be permanently removed from Hyper's systems.
We do not retain backups of deleted data beyond our standard backup rotation window.
Hyper requests the minimum OAuth scopes needed to make your connected data searchable for your agents. Here is exactly what each Google scope is used for:
| Scope | Why Hyper needs it |
| ------------------- | ---------------------------------------------------------------------------------------------------------------------------------------------------- |
| `gmail.readonly` | Read email content so agents can recall relevant message history |
| `gmail.modify` | Apply labels or filters to messages during sync — required by Gmail's API for certain organizational operations; Hyper does not delete or send email |
| `drive.readonly` | Read file content from Google Drive so agents can reference documents |
| `calendar.readonly` | Read calendar events so agents can be aware of your team's schedule |
Hyper never writes to your Drive or Calendar, and never sends email on your behalf. The `gmail.modify` scope is a Gmail API requirement for applying organizational labels during sync — it does not grant Hyper the ability to delete, compose, or send messages.
When you connect an integration, Hyper stores the resulting OAuth token encrypted at rest using **AES-256-GCM**. The encryption key is managed separately from the encrypted data, so a database breach alone would not expose usable credentials. Hyper-issued bearer tokens — used by your agents to authenticate with the Hyper API — are stored as one-way hashes. This means even Hyper's own infrastructure cannot reconstruct a plaintext bearer token after it has been issued. Service-level secrets (database credentials, API keys used internally) are stored in **AWS SSM Parameter Store** and are never written to source code, configuration files, or logs.
No. Hyper does not sell, rent, or share your data with any third party for advertising, analytics, or any other commercial purpose. Data you store in Hyper — your memories, your connected content, your agent activity — is used exclusively to operate and improve your workspace's memory layer. The only external parties that process your data are the infrastructure providers Hyper runs on (AWS), and they do so under strict data processing agreements.