These Enovari Cloud AI Terms ("AI Terms") supplement and are incorporated into the Enovari Cloud Terms of Service ("ToS"). Capitalized terms not defined here have the meanings given in the ToS. These AI Terms govern the AI-specific aspects of the Enovari platform, including Tapestry memory synthesis, the persona system, the Consensus Engine, Scanner code analysis, and knowledge graph features.
1. Definitions
- "AI Features" means Tapestry memory synthesis, the persona system, Consensus Engine ranking and retrieval, code analysis via Scanner, knowledge graph connections, and any other feature of the services that uses automated processing to generate, rank, or organize content.
- "Input" means content you provide to the services, including code, notes, queries, persona interactions, datasets, and any other material you submit for processing.
- "Output" means AI-generated content produced by the services, including summaries, knowledge connections, assessments, persona responses, code analysis results, confidence scores, and recommendations.
- "Persona Memory" means content generated by the AI persona system and stored in mind.db (private persona memory) or tapestry.db (shared knowledge memory).
- "Organizational Knowledge" means the accumulated body of Input and Output stored in your databases, including Tapestry knowledge notes, persona memories, session logs, and code analysis results.
- "Consensus Engine" means Tapestry's automated ranking and retrieval system that surfaces memories based on source credibility, evidence weight, temporal decay, access reinforcement, contradiction analysis, and other signals.
2. Ownership of AI Output
All Output derived from your Input is treated as "Your Content" under the ToS. This means:
- Silicon Harbor does not claim ownership of Output generated for you.
- You may use, modify, export, and delete Output without restriction.
- Silicon Harbor does not use your Input or Output to train, fine-tune, or improve models serving other users.
- Output is included in data export upon account termination.
The underlying AI technology, models, algorithms, system prompts, and methods that produced the Output remain the exclusive intellectual property of Silicon Harbor Technologies. The distinction is clear: you own the data and the results; we own the engine.
Copyright Notice: Under current United States copyright law (as of 2025), purely AI-generated content may not be eligible for copyright protection. Silicon Harbor makes no representation about the copyrightability of Output. Content that you author, substantially direct, or meaningfully edit retains full copyright protection. We recommend that you review the current guidance from the U.S. Copyright Office regarding AI-generated works if copyright protection is important to your use case.
3. Persona Memory and Continuity
The Tapestry persona system creates and maintains AI personas with persistent memory across sessions. The following terms govern persona memory:
- Ownership: Persona memories (stored in mind.db) are your data. They are treated as "Your Content" under the ToS.
- Privacy: mind.db is stored privately within your isolated user space. Silicon Harbor does not access, read, or review persona memory contents.
- Control: You may view, edit, export, and delete individual memories or entire persona memory databases at any time through the services interface or by request.
- Autonomous observations: Personas may generate observations and assessments based on your interactions without explicit instruction. These observations are part of the persona's contextual awareness function and are stored in your mind.db.
- Continuity: Persona continuity (session state, identity persistence, drift correction) is maintained on a best-effort basis. Continuity is not guaranteed across service interruptions, infrastructure changes, or AI model updates.
- Deletion: When you delete a persona, its associated mind.db is permanently deleted. When your account is terminated, all persona memory databases are deleted after the 30-day export window.
You acknowledge that persona responses are AI-generated and do not represent the views, opinions, or knowledge of Silicon Harbor Technologies, LLC.
4. Accuracy and Reliance Disclaimers
AI Features are informational tools. They have inherent limitations that you should understand:
- Tapestry memories may contain errors, outdated information, or AI hallucinations that do not reflect reality.
- The Consensus Engine ranks memories by confidence and other signals but does not verify factual accuracy.
- Scanner code analysis identifies potential issues but may produce false positives or miss actual vulnerabilities.
- Persona assessments reflect AI-generated analysis, not expert professional judgment.
- Knowledge connections and summaries are generated automatically and may be incomplete, misleading, or incorrect.
You should not rely on AI Output as the sole basis for business, legal, financial, medical, security, or safety-critical decisions. AI Features are supplements to human judgment, not replacements for it. Silicon Harbor is not liable for losses arising from reliance on AI-generated Output.
5. Automated Decision-Making (Consensus Engine)
The Consensus Engine is an automated decision-making system that affects which information your AI assistant sees and acts upon. Specifically, it:
- Automatically ranks and surfaces memories based on 15+ signals including source credibility, evidence weight, temporal decay, and access reinforcement.
- Assigns and adjusts confidence scores that affect how prominently information is presented.
- Detects contradictions between memories and may flag or de-prioritize conflicting information.
- Selects which persona memories to inject into AI context, affecting the AI's behavior and responses.
The Consensus Engine is an information retrieval and ranking system. It does not make decisions that produce legal effects or similarly significant effects on you as contemplated by GDPR Article 22. However, we provide the following transparency and control measures:
- Transparency: You can view the confidence scores, source attribution, and ranking signals for any memory through the services interface.
- Override: You can pin, suppress, or manually adjust the confidence of any memory, overriding the automated ranking.
- Direct instruction: You can always provide direct instructions to your AI assistant that take precedence over the Consensus Engine's recommendations.
- Human review: You may request an explanation of how a specific ranking decision was made by contacting us.
6. Organizational Knowledge and Trade Secrets
We acknowledge that your Organizational Knowledge may include trade secrets, confidential business information, strategic plans, personnel assessments, and other sensitive material. Silicon Harbor commits to the following custodianship standards:
- We treat all user content as confidential, with a duty of care consistent with the nature of the information.
- We do not review, analyze, or access your content for any purpose other than providing the services.
- We do not disclose your content to third parties except as required by the operation of the services (see sub-processor list in our Privacy Policy).
- In the event of a legal demand for your data (subpoena, court order, regulatory request), we will notify you before disclosure unless prohibited by law or court order.
We recommend that you implement your own content classification policies and exercise judgment about what information to store in cloud-based AI memory systems. Do not store data in Enovari that you would not entrust to a cloud storage provider with equivalent security measures.
7. AI Model Provider Transparency
Certain AI Features may involve processing by third-party AI model providers. We are transparent about this:
- The Enovari platform's core analysis, parsing, and retrieval systems are proprietary technology built by Silicon Harbor.
- Your MCP-connected AI client (the AI assistant you use to access Enovari, such as Claude, ChatGPT, Gemini, etc.) is a third-party service not operated by Silicon Harbor. Your interactions with that client are governed by that provider's terms.
- If Silicon Harbor integrates additional third-party AI processing in the future, we will disclose the provider, the data processed, and the purpose before the integration takes effect.
8. Modifications to AI Features
We may update, modify, or discontinue specific AI Features. We commit to the following:
- Material changes to how Output is generated or how your data is processed will be communicated via email and in-app notice at least 30 days before they take effect.
- Changes to the AI model providers or processing methods underlying AI Features will be disclosed in advance.
- If you find material changes unacceptable, you may export your data and terminate your account in accordance with the ToS.
9. Contact
If you have questions about these AI Terms, the Consensus Engine, persona memory, or how AI Features process your data, please contact us:
Silicon Harbor Technologies, LLC
Charleston, South Carolina, United States
Email: legal@silicon-harbor.net