Capable and contained.
Azros is built on infrastructure with model gating, content moderation, and frontend protection woven through every surface.
Model gating
The most capable models are gated by plan, not just priced. The frontier trio and realtime voice unlock on Premium and above, mid-tier models from Growth, while the free default and core trio serve every account. The full map lives on the pricing page.
Moderation
Prompts and generations pass through layered content filtering tuned per modality, with jailbreak detection on text and stricter thresholds on image and voice output. Accounts that push abuse boundaries are rate-limited, then suspended under the usage policy.
Shield
What you build gets protected too. Shield compiles frontend JavaScript into WASM with AES-256-GCM encryption, ephemeral session keys, and domain-whitelisted execution. See Shield on the capabilities page.
Data handling
Data is encrypted in transit and at rest. Conversation history persists only for your own recall and can be deleted from the Console at any time; deleted accounts and their memories are purged within the retention windows published in the privacy policy.
Age gates
Accounts require a minimum age of 18, and unrestricted model access sits behind explicit account-level verification. Voice and storefront features follow the same gate, so what a minor can reach is what a minor should reach.
Reporting abuse
Seen something that should not exist on the platform - a scam storefront, abusive generation, or misuse of the affiliate chain? Report it at the contact page and it is triaged the same day.
Read the full policies.
The complete legal and policy library is published and maintained by Agentics, the company behind Azros.