Launching today
kosa8
Air-gapped microVM sandboxes for enterprise AI agents
135 followers
Air-gapped microVM sandboxes for enterprise AI agents
135 followers
kosa8 is security infrastructure for enterprise internal AI agents. Every agent gets its own microVM with its own kernel, on your Macs, your Linux servers, your cloud, or fully air-gapped. The agent sees one folder, API keys stay on the host (the sandbox holds a placeholder), and the network is closed except the hosts you allow. Rewind the whole machine in about a second when an agent gets it wrong. Docker-compatible, MCP-native, hash-chained audit, signed releases.
Free Options
Launch Team / Built With
AWS Startup AdvisorPersonalized AWS guidance that meets you where you’re already building
Promoted
@jesv Human oversight: signed, two-person approvals
@Justin Jincaid Kosa8 comparison with all checklists
@Shivam Kushwaha kosa8 puts each agent in its own sealed microVM. It sees only one folder, the network is closed except for sites you allow, and real API keys never go inside. All of this is enforced from outside the VM, so the agent can't switch it off. Risky actions wait for a human to approve them. Any mistake can be rewound in about half a second. And every attempt is recorded in a log nobody can quietly edit.
@hamza_afzal_butt Building a zero‑trust pipeline with sealed microVMs
When you combine sealed microVMs, an egress broker, and an immutable audit trail, you create a zero‑trust environment for AI agents. The agent never holds secrets, cannot reach the internet except through approved hosts, and any deviation is captured for forensic review.
If you need to test alternative approaches, fork the running machine into live copies—up to eight at once—so you can experiment in parallel without contaminating the original state. Once the winning variant is identified, delete the failed forks and keep the clean checkpoint.
@Swati Raghuwanshi A prompt telling an agent to behave is a suggestion, while a sealed machine it can't escape is a rule. That's why kosa8 gives every agent its own microVM, keeps real keys outside it, and logs every action.
yes @fmerian AI agents now run code, touch files and call APIs, but most regulated teams (banks, defense suppliers, healthcare) can't let them do that without containment and proof.
kosa8 gives them both:
• A sealed machine for every agent. Each agent runs in its own microVM with its own kernel and sees only one workspace folder.
• No secrets inside. Real API keys stay on the host, and network access is limited to an allow-list, both enforced outside the VM.
• Humans in control. Named approvers can be required for risky actions, and rewind restores a running machine in about half a second.
• Proof for auditors. A tamper-evident, hash-chained log records every action, and FIPS 140-3 builds are available for government work.
• Runs where the data lives. It works on-prem, air-gapped or on the customer's own cloud, with local models, so nothing leaves the building.
@kevinKosa8 about Letting your AI agents do real work, safely. kosa8 gives every agent its own sealed workspace: a private microVM that sees only the folder it needs. API keys stay safely on your host, the network reaches only the sites you choose, and important actions wait for a person's approval. Every step is recorded in a tamper-evident log, so you always know exactly what happened, and the whole workspace can be rewound in half a second. It runs on your laptop, your servers or fully offline, so your data always stays with you. Your team gets the speed of AI agents, with control and peace of mind built in.