Platform / Agent Infrastructure
A real machine for every agent.
Available
Coding agents need more than a container and a prompt. OhMako gives each agent its own Firecracker machine with repositories, dependencies, tools, and scoped secrets — reachable over shell, SSH, MCP, and CLI.m-4c19 · agent machine
$ ohmako machine create --env acme-monorepo --from snap_8e21
Environment as configuration
Define repositories, toolchains, services, and secrets once. Every agent machine starts from the same definition.
Sleep, don't destroy
Idle machines sleep and wake with their state intact, so long-running agent work survives pauses.
Humans can step in
Attach a terminal, SSH session, or browser view to see exactly what the agent sees.
Capabilities
What's in Agent Infrastructure
- Firecracker-based isolated machines
- Multi-repository environments
- Configurable environments and secrets
- Shell, SSH, terminal, and file access
- Browser and VNC access where supported
- MCP and CLI access
- Sleep and wake
- Machine snapshots and restoration
- Persistent development environments
Connected
How it fits the rest of OhMako.
AGT → SNP
Snapshots and Warm Pools
Agent machines start from snapshots in seconds instead of rebuilding.
AGT → REVAI Code Review
Review findings are dispatched to agents for remediation.
AGT → MRGMerge Queue
Agent pull requests enter the same validated merge path.
Works with
MCP clientsCoding agentsGitHubGitLab
Build the next generation of software delivery.
Give your developers and AI agents the infrastructure to build, validate, and ship software without the operational friction.