Everything you need to choose your crew.

A nanion is a persistent, named AI entity you specialize and put to work. It has a clear core, operating rules, tools, scoped permissions, memory, and a work history. Unlike a disposable chat persona, its role and relationship persist.

Nanions is a native command center for putting persistent AI entities to work across your files, apps, and services. Give your Chief of Staff an outcome, follow the job, approve sensitive actions, and receive the result with its sources and receipts.

No. You are the Commander-in-Chief, or CinC. Your Chief of Staff nanion serves as your second-in-command and main point of contact. It appoints Crew Leaders, who direct as many Executors as needed. You set direction and receive the decisions that need your authority.

They share the same product foundation and persistent AI entities. Nanions is tuned for research, strategy, operations, documents, content, personal systems, and repeat work. NanX, short for Nanions eXecutor, gives software work a code-focused command center with repositories, terminals, worktrees, diffs, tests, and review.

A loop is recurring work that still needs judgment, such as triage, monitoring, research, or a daily brief. A workflow codifies repeat work that needs fixed logic: stages, branches, schedules, retries, checkpoints, and approval gates. A workflow can call a nanion only where AI judgment helps.

Nanions are designed to act with the tools and permissions you grant. New or sensitive actions can require confirmation. Trust may expand through successful work, but destructive or consequential actions can remain gated. You can interrupt or wind down work at any time.

Nanions supports a choice of intelligence: supported provider accounts, local models, and managed Omni. Your Chief of Staff can use the right allowed model for a job without changing the nanion’s identity, tools, memory, or role.

Raw memory stays on your devices. Derived memory and context sync through your secure Nanion Cloud so your crew can continue across your approved devices. When a job uses a connected model provider or service, Nanions sends only the derived context selected for that job.

Yes, through approved tools and connections. Access is scoped by nanion and requested when needed. Connecting an account does not give every member of your crew blanket access.

Nanions launches on macOS first. Windows and Linux follow once their installation, permissions, updates, and recovery meet the same quality bar.

Personal costs $5 per month with your own supported inference. Plus costs $30 per month and includes managed Omni credits. Crew and Company plans cost $100 and $200 per month. Plans go on sale when downloads open.

Yes. Start with first-party roles such as Jarvis, Friday, or Strategist, adapt a specialist, or build a custom nanion around your own work, tools, rules, and preferred way of collaborating.