Privacy Policy

01. Our Stance on Privacy

Prism AI is designed with an offline-first, client-authoritative privacy paradigm. As an orchestration framework built under a Bring Your Own Credentials (BYOC) pattern, we fundamentally eliminate standard server-side intermediation. We do not persist, compile, or relay your secure credentials, API tokens, or code environments to third-party databases.

02. Information We Collect

Because Prism AI executes code sandboxing and prompt-handling client-side, we preserve your privacy by severely minimizing standard payload ingestion.

• API Keys & Tokens: Extracted directly from your custom configuration environment. They remain bound strictly to volatile memory or local browser environments, never transmitted back to Prism AI Labs servers.
• Usage Metrics: Fully anonymous diagnostics to optimize sandboxing frame rates and code compile successes.
• Early Access Directory: When you voluntarily submit your application to the active build list, we preserve your primary input parameters (such as email and context answers) to handle admission lists safely.

03. Secure Sandbox Execution

The execution matrices utilized in standard Prism AI actions are local or deeply isolated processes using local VM layers. There is no background data tunneling. Any diagnostic output generated from user-loaded instructions is deleted immediately upon closure of your browser window.

04. Data Retention and Safety

Under no circumstances do we trade, distribute, or broker your organizational parameters or LLM inputs. Your localized credential inputs are retained exclusively in standard storage configurations inside your sandbox environments (with no outbound replication hooks).