kernel infrastructure · open source

virtual machines that feel like containers.

run oci workloads in isolated vms across cloud hypervisor, firecracker, qemu, and apple virtualization.framework.

curl -fsSL https://get.hypeman.sh | bash
tl;dr

install hypeman with one command, then use the docker-like cli or typed sdks.

58api operations
77openapi schemas
2generated sdks
4hypervisors
01 / docs

start simple. go deep when you need to.

the shortest path to a running vm, plus focused guides for the lifecycle, networking, storage, and host configuration.

01

install hypeman

server, cli, token tool, generated configuration, and system service in one installer.

read installation guide →
02

vm lifecycle

create, inspect, stop, standby, restore, snapshot, and fork isolated workloads.

explore lifecycle →
03

remote access

connect a separate cli or sdk with a server url and jwt-backed api key.

configure a client →
04

ingress + tls

route hostnames to guest ports with reverse proxying, tls termination, and subdomains.

configure ingress →
05

host configuration

select a hypervisor and tune networking, storage, images, snapshots, gpu, and logging.

open linux config →
06

contribute

build locally, run tests, understand services, and work across supported host platforms.

read development guide →
quickstart

from zero to a running vm.

the installer configures the local server and cli. pull any oci image, run it, then exec directly inside the guest.

# install server + cli
curl -fsSL https://get.hypeman.sh | bash
# boot an oci workload
hypeman pull nginx:alpine
hypeman run --name my-app nginx:alpine
# inspect and enter the vm
hypeman ps
hypeman exec -it my-app /bin/sh
02 / api reference

the complete http surface.

generated from openapi.yaml at kernel/hypeman@4b55ba9a. expand a resource to scan methods, paths, and operation summaries.

10 resources · 58 operations

health 1 endpoint
methodpathdescription
GET /health Health check
host resources 2 endpoints
methodpathdescription
GET /resources Get host resource capacity and allocations
POST /resources/memory/reclaim Trigger proactive guest memory reclaim
images 4 endpoints
methodpathdescription
GET /images List images
POST /images Pull and convert OCI image
GET /images/{name} Get image details
DELETE /images/{name} Delete image
instances 23 endpoints
methodpathdescription
GET /instances List instances
POST /instances Create and start instance
GET /instances/{id} Get instance details
PATCH /instances/{id} Update instance properties
DELETE /instances/{id} Stop and delete instance
POST /instances/{id}/standby Put instance in standby (pause, snapshot, delete VMM)
POST /instances/{id}/restore Restore instance from standby
POST /instances/{id}/fork Fork an instance from stopped, standby, or running (with from_running=true)
POST /instances/{id}/snapshots Create a snapshot for an instance
GET /instances/{id}/snapshot-schedule Get snapshot schedule for an instance
PUT /instances/{id}/snapshot-schedule Create or update snapshot schedule for an instance
DELETE /instances/{id}/snapshot-schedule Delete snapshot schedule for an instance
POST /instances/{id}/snapshots/{snapshotId}/restore Restore an instance from a snapshot in-place
POST /instances/{id}/stop Stop instance (graceful shutdown)
POST /instances/{id}/start Start a stopped instance
GET /instances/{id}/logs Stream instance logs (SSE)
GET /instances/{id}/stat Get filesystem path info
GET /instances/{id}/wait Wait for instance to reach a target state
GET /instances/{id}/auto-standby/status Get auto-standby diagnostic status
POST /instances/{id}/auto-standby/hold Hold off auto-standby
GET /instances/{id}/stats Get instance resource utilization stats
POST /instances/{id}/volumes/{volumeId} Attach volume to instance
DELETE /instances/{id}/volumes/{volumeId} Detach volume from instance
snapshots 4 endpoints
methodpathdescription
GET /snapshots List snapshots
GET /snapshots/{snapshotId} Get snapshot details
DELETE /snapshots/{snapshotId} Delete a snapshot
POST /snapshots/{snapshotId}/fork Fork a new instance from a snapshot
volumes 5 endpoints
methodpathdescription
GET /volumes List volumes
POST /volumes Create empty volume
POST /volumes/from-archive Create volume from archive
GET /volumes/{id} Get volume details
DELETE /volumes/{id} Delete volume
devices 5 endpoints
methodpathdescription
GET /devices List registered devices
POST /devices Register a device for passthrough
GET /devices/available Discover passthrough-capable devices on host
GET /devices/{id} Get device details
DELETE /devices/{id} Unregister device
ingresses 4 endpoints
methodpathdescription
GET /ingresses List ingresses
POST /ingresses Create ingress
GET /ingresses/{id} Get ingress details
DELETE /ingresses/{id} Delete ingress
builds 5 endpoints
methodpathdescription
GET /builds List builds
POST /builds Create a new build
GET /builds/{id} Get build details
DELETE /builds/{id} Cancel build
GET /builds/{id}/events Stream build events (SSE)
builders 5 endpoints
methodpathdescription
GET /builders List builders
POST /builders Create builder
GET /builders/{id} Get builder details
DELETE /builders/{id} Delete builder
POST /builders/{id}/prune Prune builder cache

machine-readable source. download the raw openapi document or review it on github. request and response schemas are generated into the sdk references below.

03 / sdk reference

typed clients, generated with the api.

official TypeScript and Go libraries expose the same resources with typed request fields, response objects, retries, timeouts, and structured errors.

TypeScriptv0.5.053 methods

@onkernel/hypeman

10 generated resource groups with editor-ready request and response types for server-side TypeScript and JavaScript.

npm install @onkernel/hypeman
Gov0.22.053 methods

github.com/kernel/hypeman-go

11 generated resource groups with typed params, context-aware calls, request options, retries, and standard Go errors.

go get github.com/kernel/hypeman-go@v0.22.0