Safety and trust
Control should stay on your side.
SpectraMind AI is built to clearly separate personal memory, project knowledge and temporary conversations — and to never present uncertain information as confirmed.
Transparency
SpectraMind is designed to avoid inventing facts or presenting uncertain information as confirmed. When the available official information isn't enough, that should be said honestly.
Privacy and control
Normal conversations, temporary chats, personal memory and project knowledge are kept separate — and the user can review and manage their personal memories.
Data treated as data
Information retrieved from memories, projects or external sources is treated as data, not as system instructions.
Authorization for integrations
Features connected to external accounts — such as Excel, Gmail or Roblox Studio — require proper authorization before they act.
What this means in a normal conversation.
Personal memory
Facts and preferences useful for future conversations. The user is the one who reviews, corrects and deletes what's been saved.
Temporary chat
Doesn't appear in history and doesn't create or use memories — built for conversations that shouldn't leave a trace.
Project knowledge
Stays tied to the project where it was shared and remains separate from the user's personal memory.
New capabilities — including agents, connected tools and models still in preview like SpectraMind PRO v2 — must come with controls, clear limits and honest communication about what is, or isn't, available yet.
Read the privacy policy.
For the complete legal terms on data, retention and user rights.