AI Model Safety & Trust

Trust is the foundation of Novellis. We provide tools that ensure your intellectual property is protected and that your AI interactions are secure.

1. No Training on Your Data

When you connect to established providers like OpenAI, Anthropic, or Google, we utilize their API platforms.

• Zero Retention: We configure our API calls to opt-out of data training by default where the provider allows.
• Direct Connection: Your API keys are stored locally on your machine and are only used for direct communication with the model provider.

2. Local Models (Ollama)

For absolute certainty, we recommend using Ollama or other local inference engines.

• completely Air-Gapped: You can run Novellis and your AI model entirely offline.
• No Third-Party Access: Your lore and drafts never touch any server, including ours.

3. Transparency

Novellis provides a Heads-Up Display (HUD) for all AI interactions:

• Token Usage: See exactly how much each request costs.
• Context Awareness: Know exactly which files were sent to the AI for each prompt.
• Audit Logs: Review past AI interactions to see what was discussed and what data was referenced.

4. Your IP remains Yours

The content generated by AI in Novellis is your property. We do not claim any rights to your manuscripts or world-building notes.