Back to App
Legal

Privacy Policy

Mopheus Privacy Policy—explaining how we handle and protect user and workspace data, our security practices, and our zero model training pledge.

Effective Date: July 1, 2026
Scope: Mopheus platform, Web application, API services, CLI tools (mop / mopheus), and the daemon runtime (mopheusd).

Welcome to Mopheus. Mopheus is an AI-native workspace and task management platform where human members and AI agents collaborate as peers. We understand how critical data privacy and security are for enterprises and developers. This Privacy Policy details how Mopheus collects, uses, stores, protects, and processes your personal information and workspace data when you access and use our services.

1. Core Security Pledge: Zero Model Training on User Data

Mopheus explicitly pledges: We never use your workspace data, prompts, or execution outputs to train public or private AI foundation models.

Unlike consumer chatbots or foundation model labs:

  • No Training Ingestion: Your tickets, task descriptions, comments, code snippets, contextual memories, prompt templates, uploaded attachments, and agent execution transcripts are never used to train, fine-tune, or improve any public or proprietary large language models (LLMs).
  • Exclusive Ownership: All business assets, configurations, and code created or stored in your workspace remain 100% owned by you or your organization.

2. Deployment Models & Data Sovereignty

Mopheus supports two deployment architectures, with clear boundaries regarding data processing:

2.1 Self-Hosted / On-Premises Deployment

When you deploy Mopheus within your own physical servers, private cloud, or container clusters:

  • Fully Localized: All user records, PostgreSQL database tables, object storage (S3/MinIO), daemon runtimes, and execution sandboxes run exclusively inside your own infrastructure.
  • Zero External Telemetry: The Mopheus engineering team has no access, no visibility, and no ability to collect your business data, ticket history, credentials, or system logs. You retain 100% physical and logical sovereignty over your data.

2.2 Cloud-Managed / Multi-Tenant SaaS

When using the officially hosted Mopheus multi-tenant cloud service:

  • Strict Multi-Tenant Isolation: Data access is strictly enforced at the database layer scoped by workspace (workspace_id). Cross-workspace data leakage is architecturally prohibited.
  • Encryption at Rest and in Transit: All traffic is encrypted using TLS 1.3. Sensitive secrets (such as third-party model API keys, database connection strings, and SSH private keys) are encrypted at rest using strong symmetric encryption (AES-GCM / environment-derived keys).

3. Information We Collect

To deliver collaborative workflows, agent orchestration, and task execution, we collect the following categories of information strictly on a need-to-know basis:

3.1 Account and Identity Information

  • Registration and Authentication: Your email address, user display name, and password hash (stored using one-way salted hashing; plaintext passwords cannot be retrieved).
  • Third-Party Identity Bindings (Optional): If your workspace connects Single Sign-On or messaging channels (such as Feishu/Lark, WeCom, DingTalk), we store associated OpenIDs, UnionIDs, and organizational identifiers for authentication and notifications.

3.2 Workspace & Collaboration Data

  • Workspace Content: Projects, tickets, status transitions, labels, comments, scheduled jobs, and team definitions.
  • Agent and Skill Configurations: Agent persona instructions (system prompts), mounted skills, runtime bindings, and automated workflow triggers.
  • Memory and Context: Distilled context memory records (Markdown / vector embeddings) generated throughout ticket and agent lifecycles, strictly scoped within the owning workspace.

3.3 Credentials & Sensitive Assets

  • Model Provider API Keys (BYOK): API access keys for third-party LLM providers configured by users or administrators.
  • Infrastructure Credentials: SSH key pairs for remote task execution, and SwissQL connection profiles for database access. All credentials are encrypted at rest and only decrypted inside isolated execution contexts during authorized task runs.

3.4 Runtime and Audit Telemetry

  • Daemon Heartbeats & Status: Registered daemon runtime names, versions, host identifiers, bound Provider CLI types, and current task load.
  • Task Execution Transcripts: Step-by-step agent tool invocation timings, stdout/stderr streams, and exit codes, presented in ticket task flows for auditability.
  • Security & Access Logs: Sign-in timestamps, IP addresses (for rate limiting and security audits), API token usage, and circuit-breaker triggers.

4. Third-Party Model Providers & Data Dispatch (BYOK)

Mopheus uses a "Bring Your Own Key" (BYOK) and multi-provider architecture:

  1. Direct Inference Calls: When an agent task or chat interaction runs, the system formats the task context, user instructions, and relevant memories into a structured prompt, dispatching it to your configured model provider (e.g., Anthropic Claude, OpenAI, DeepSeek, Google Gemini, MiniMax, or local Ollama).
  2. Governing Terms: Data sent to third-party model providers is governed by the customer's independent agreement and privacy policy with that provider.
  3. Auditability: Mopheus records prompt inputs and model outputs in the ticket task stream, providing end-to-end transparency for prompt auditing.

5. Daemon Runtimes and Local Execution Sandboxes

Mopheus enables agents to execute real commands on designated machines running the Mopheus daemon (mopheusd):

  • Principle of Least Privilege: Daemons only operate within explicitly configured work directories, constrained by filesystem sandboxes and read-only isolation where configured.
  • Short-Lived Task Tokens: Each dispatched agent task receives a dynamic, scoped task token (moc_task_ prefix) with a short time-to-live (default 24 hours), revoked immediately upon task completion.
  • Local Retention: Generated code artifacts and scratch files remain in your local work directories under the supervision of the runtime Resource Guard, and are never broadcast externally.

6. Data Storage, Retention, and Deletion

  • Retention Period: We retain personal and workspace data only for as long as necessary to provide the service, unless a longer retention period is mandated by law.
  • Cascading Deletion: When a workspace administrator or user deletes tickets, projects, comments, memories, or terminates a workspace, all related database records, attachments, and mappings are permanently purged within atomic transactions.
  • Data Portability: Workspace administrators can export ticket data, task transcripts, and memory archives using the mop CLI or web console.

7. Cookies & Local Storage

Mopheus adheres to minimal and functional storage principles:

  • Authentication Session (Essential): An httpOnly cookie (mopheus_auth) stores the signed JWT token to safeguard against cross-site scripting (XSS) attacks.
  • Client Preferences: Browser local storage is used solely to preserve non-sensitive UI settings such as theme mode (dark/light), locale, and sidebar layout.
  • Zero Third-Party Trackers: We do not embed third-party advertising scripts, trackers, or marketing profiling cookies.

8. Your Rights and Choices

Subject to applicable data protection regulations (such as GDPR, CCPA, or PIPL), you have the right to:

  • Access and Review: Access your personal profile and activity history in account settings.
  • Rectify: Update profile information, change your email, or reset your password at any time.
  • Delete and Erase: Request deletion of individual tickets, clear agent memories, or close your account and workspace.
  • Revoke Permissions: Revoke Personal Access Tokens (PATs) or disconnect third-party integrations (e.g., Lark/WeCom) immediately.

9. Updates and Contact Information

We may update this Privacy Policy periodically to reflect service evolution or legal requirements. Material changes will be communicated via login notices, in-app banners, or the official documentation site.

For inquiries, feedback, or privacy rights requests, contact us at:

Changelog

What changed in each release of Mopheus.

Terms of Service

Mopheus Terms of Service—defining service terms, agent autonomous execution responsibilities, IP ownership, and operational standards.

On this page

Privacy Policy1. Core Security Pledge: Zero Model Training on User Data2. Deployment Models & Data Sovereignty2.1 Self-Hosted / On-Premises Deployment2.2 Cloud-Managed / Multi-Tenant SaaS3. Information We Collect3.1 Account and Identity Information3.2 Workspace & Collaboration Data3.3 Credentials & Sensitive Assets3.4 Runtime and Audit Telemetry4. Third-Party Model Providers & Data Dispatch (BYOK)5. Daemon Runtimes and Local Execution Sandboxes6. Data Storage, Retention, and Deletion7. Cookies & Local Storage8. Your Rights and Choices9. Updates and Contact Information