Introducing the Vantisso Trial

We're releasing the first Trial of Vantisso. Vantisso is a security- and control-focused agent execution platform that runs each AI agent isolated inside a Firecracker MicroVM and reclaims it completely when the work is done. This Trial lets you verify it through hands-on use — from install to your first agent to running a team of them.

What the Trial ships

• Hardware-isolated execution — a Firecracker MicroVM spins up in milliseconds per agent task and is torn down without a trace afterward, disk and network included. The result is a clean execution environment with no cross-contamination between runs.
• A built-in agent runtime — a native Go runtime inside each VM is built to use the APIs of all four providers available today: google · anthropic · openai · groq.
• A console built for two personas — an Operator plane for infrastructure (VMs, network, storage, security) and a Workspace plane for composing agent teams and running collaboration, serving two distinct personas.
• From a single agent to agent teams — spawn one agent, or orchestrate a team of them collaborating across roles (broadcast, a shared feed, and handoff).
• Security built in — hardened security is on by default: an egress firewall that governs each VM's outbound traffic (blocking the cloud metadata range to prevent credential theft) and source IP/MAC anti-spoofing.
• Fast to start — snapshot cold restore in under 200ms, and spawns in as little as 2ms with a warm pool.
• An integration REST API — a well-structured REST API lets you control and manage Vantisso easily from your own application or your company's IT operations portal.

Also included: the MCP gateway, snapshots, durable chat transcripts, and token streaming.

30 days, every feature, right after install

The point of this Trial is simple: you can actually run it.

• Install immediately — it ships as prebuilt packages (FULL / SLIM), so it installs straight onto a Linux host with no Go toolchain or source checkout.
• A full 30-day evaluation — each package includes an evaluation license valid for 30 days from first run, unlocking every feature with no key-issuing step.
• Your first agent in clicks — create a VM in the web console and chat with the agent right in the conversation panel. No wrestling with the CLI — you can experience everything Vantisso does through the Web UI alone.
• After the evaluation ends — reading existing agents, running tasks, and tearing them down keep working; only spawning new VMs or teams stops. The console's top banner always shows the time remaining.

# Grab a prebuilt package and install (see Getting Started for requirements and steps)
tar xzf vantisso-eval01-linux-amd64-full.tar.gz
cd vantisso-eval01
sudo ./install.sh

# Once install finishes, open the console in your browser and run your first agent
#   http://localhost:3000/ui/

Getting started

Requirements (Linux · KVM · root) and step-by-step instructions live in Getting Started; for integrating your own application, use the API Reference.

For a commercial license or an extended evaluation, reach us at sales@ax-onn.com.