Privacy Policy

At Prema AI Labs, we believe that intelligence should be sovereign. Your conversational data, business strategy logic, and intent graphs belong entirely to you. Our architecture is designed to enforce this mathematically.

1. Data Sovereignty & Inference

We do not retain, store, or warehouse any conversational logs generated through the local-inference options of the Prema.ai agent platform. For our cloud-hosted endpoints, telemetry and prompt inputs are retained strictly in volatile memory for the duration of the API cycle to generate your response, after which they are instantly overwritten.

2. Model Training Protocol

We will never train our foundational models on your proprietary or personal data. Any fine-tuning, RAG (Retrieval-Augmented Generation), or contextual memory graphs created by our software live exclusively within your authenticated hardware environments or isolated, zero-trust cloud VPCs. We explicitly opt-out all customer endpoints from synthetic data generation programs.

3. Data Collection

The only data we collect belongs strictly to user-level analytics necessary to diagnose crashes and compute hardware utilization. This includes:

• System architecture specs (RAM, GPU presence) to optimize the WebGL and rendering engine performance.
• Anonymized routing errors (404 and 500 status codes).
• Account billing information, handled securely by our heavily audited financial partners (Stripe).

4. Your Rights

In accordance with GDPR, CCPA, and global best practices, you possess the absolute right to demand the erasure of any remnant structural data associated with your account. Due to our decentralized architecture, terminating your account via your settings dashboard immediately and irreversibly cascades a cryptographic deletion of your identity footprint across all our ledgers.