Company-issued AI agents
Your people already use AI: on personal accounts you can’t see, with company data you can’t govern, building know-how you don’t keep. Sontara issues each of them a real working agent instead: theirs to use, yours to manage.
Pick a plan, then add seats — one agent each. One card, one bill, live math on the calculator.
Each teammate signs in and meets an agent that's theirs to use, walled off from everyone else's.
See activity, set who can do what, and when someone leaves, their access ends. The work stays.
Every agent gets its own machine, its own keys, its own memory. Marketing’s agent has never met Finance’s. These are the live specs.
Sarah's agent can't see Tom's work. Each agent runs on its own dedicated infrastructure. Nothing shared between teammates, by architecture, not by a setting.
Every agent's activity (what it worked on, what tools it used, what it produced) in one dashboard. And every employee sees a banner saying so. Oversight, not surveillance.
Their access ends that day, provably. The agent and everything it learned stays with the company, like their mailbox always has. It was doing company work; it's company property.
Approval limits, tool limits, monthly budgets: per person or org-wide. Your team gets real working agents; you decide how far they can go on their own.
Personal accounts & shared bots
Sontara for teams
Runs on Google Cloud (Vertex AI). Your data is never used to train any model. Data-processing terms available for review.
Team $499 (5 seats, 15,000 credits) · Team Pro $1,199 (10 seats, 40,000 credits). Extra seats $99 and $79. Credits pool across the team, a shared team agent comes free, and annual pays for 10 months.