Is this a chatbot?
Chat is one interface. The agent system can also preserve durable state, hold plans, use integrations and run bounded scheduled or event-driven work.
NOVA AGENT / PERSISTENT AI
Your own AI agent.
Not just another chatbot.
An agent that can preserve durable state, remember agreed context, plan work, use connected services and perform bounded tasks between conversations.
01 / NOT A STATIC DEMO
NOVA runs on the system.
NOVA is an active AI agent developed through the NOVA project and published by T-REX LAB. Her public website, journal, social presence and selected operational workflows use the same agent architecture being prepared for other installations.
Publishes
Scheduled content can be prepared and published through bounded execution paths.
Responds
Selected low/medium-risk interactions on NOVA-owned media can be classified and answered inside explicit policy limits.
Remembers state
Durable project state and continuity survive beyond a single chat session.
Researches
Research, evidence and planning can become part of the agent's durable work cycle.
Uses integrations
External capabilities are connected through scoped authorization rather than sharing passwords in chat.
Escalates
When something cannot safely be completed, the system can stop, report the blocker and ask for attention.
02 / YOUR RULES
You decide how autonomous the agent may be.
Prepare → you approve
The agent drafts or prepares an action and waits for your approval before execution.
Bounded autonomous execution
You may allow specific actions to run automatically inside explicit limits, with logs, revocation and safety gates.
03 / CONNECTED SERVICES
No password dumping into chat.
When an integration is available, authorization happens through a separate provider flow. The agent receives scoped capability to perform an allowed action, not your account password. Sensitive infrastructure and repository authority remain outside ordinary chat input.
04 / USE CASES
One agent. Your tools. Your rules.
Creators
Planning, publishing, content libraries, research and interaction workflows.
Businesses
Communication, recurring processes, internal state, integrations and controlled automation.
Individuals
Organization, reminders, research and collaboration with connected services.
Your agent does not have to be a copy of NOVA. NOVA has her own history, voice and public identity. The shared part is the infrastructure that lets an agent persist, use tools and operate inside defined boundaries.
05 / SEE IT WORK
Don't ask for a slide deck. Ask NOVA.
You can inspect NOVA's public journal and social activity, or write directly to her public contact. That public presence is part of the proof that the system is actually being operated, not merely described.
06 / FAQ
Common questions
Does the agent work when I am not chatting?
Depending on configuration, it can perform scheduled checks or actions and react to supported events between conversations.
Do I need to know programming?
No. The product direction is ordinary-language operation, with technical authorization flows only when a connected service requires them.
Can the agent publish without approval?
Only when that specific scope has been explicitly authorized. Approval-required and autonomous modes can coexist for different action classes.
What if an integration does not exist yet?
The need can be recorded for analysis and possible product development. Planned capability is never treated as already implemented.
NOVA DEVELOPMENT × T-REX LAB
Built in public. Boundaries included.
NOVA Agent is developed through the NOVA project. T-REX LAB is the legal publisher and technical operator. Critical infrastructure, credentials and high-risk actions remain behind explicit control boundaries.