Skip to content

Beta: these docs are still being updated. The next full update is on 18 November at 8 PM UTC. Docs in beta. Full update 18 Nov, 8 PM UTC.

Hoody.com

Lookup tables for Computer Images. Every call, field and status is in the Computer Images API.

OSVersionImage NameSize
Debian13 (Trixie)debian/13~500 MB
RuleDetail
What it is fordebian/13 is the base image for every workload: development, production services, databases, and Docker or Kubernetes hosts
What it bringsglibc and the apt package manager, with Debian’s full package repository; stable, slow-moving releases with security updates; a minimal but complete environment
SpellingsCanonical debian/13. Also accepted: debian:13, debian-13, debian 13, and debian (resolved to debian/13)
Image stays unchangedCreating a computer from an image does not change the image
WindowsThe computer’s base image is debian/13. To run Windows inside a computer with KVM enabled, use Hoody VM
CallWhat it does
GET /api/v1/images/publicBrowse available images
GET /api/v1/images/public/{id}Get image details
GET /api/v1/images/{id}/iconDownload image icon
GET /api/v1/images/userList imported/purchased images
POST /api/v1/images/import/{id}Import free image
POST /api/v1/images/purchase/{id}Purchase paid image
POST /api/v1/images/rate/{id}Rate image (0-5 stars)
PropertyDescriptionExample Values
OSOperating systemdebian
ReleaseVersion/release13
ArchitectureCPU architectureamd64
SizeDisk space requiredabout 500 MB
PriceCost in USD0 (free); a paid image shows its USD price
RatingCommunity rating0.0 - 5.0 stars
VariantSpecial configurationdefault

The details of one image (GET /api/v1/images/public/{image_id}) answer:

{
"statusCode": 200,
"message": "Computer image details retrieved successfully",
"data": {
"id": "63a3e4b5c6d7e8f9a0b1c2d3",
"alias": "debian/13",
"description": "Debian 13 (Trixie)",
"image_name": "debian/13",
"architecture": "amd64",
"os": "debian",
"release": "13",
"serial": "20250809",
"variant": "default",
"size": 524288000,
"price": 0,
"added_date": "2023-11-09T00:00:00.000Z",
"average_rating": 4.8,
"rating_count": 1247,
"icon_url": "/api/v1/images/63a3e4b5c6d7e8f9a0b1c2d3/icon"
}
}

Use the details to check the size for storage planning, read the description for pre-installed software and see community ratings for quality.

GET Browse available images
/api/v1/images/public
Click "Run" to execute the request
Query parameterFilters by
osOperating system (debian)
architectureCPU architecture (amd64, arm64, armhf)
min_price / max_pricePrice range
min_ratingMinimum rating
searchKeyword, matched against the image’s alias, name and description
sort_by / sort_orderSort order of the results
max_ratingMaximum rating; ratings run from 0 to 5
page / limitPage of results: pages start from 1, and a page holds up to 100 images (default 20). Your library (GET /api/v1/images/user) pages the same way; hoody images list --mine lists it
Terminal window
# Highest rated images
hoody images list --sort-by rating --sort-order desc --limit 10
Terminal window
# Free images only
hoody images list --min-price 0 --max-price 0
# Budget images ($0-$10)
hoody images list --min-price 0 --max-price 10
# Premium images ($25+)
hoody images list --min-price 25

An image’s architecture field (from GET /api/v1/images/public) shows which CPU it targets. A create names only debian/13, and the computer gets its server’s architecture automatically, so there is nothing to match. The values mean:

ArchitectureProcessorsNotes
amd64 (x86-64)Standard Intel and AMD processorsThe most common value
arm64 (ARM 64-bit)ARM processors, such as Apple Silicon, AWS Graviton, and Raspberry Pi 4 and laterARM-based servers and cost-optimized cloud instances
armhf (ARM hard float)Older ARM devices, Raspberry Pi 3 and earlierRarely needed in modern deployments

The variant field names an image’s configuration of its OS version. debian/13 shows default, the default full-featured configuration for general use.

Specify the image by its alias (debian/13). The variant is exposed as a separate variant field in each image’s marketplace metadata; it is not selected by appending a suffix to the alias.

RatingMeaning
5 starsExcellent (works perfectly, well-configured)
4 starsGood (minor issues or missing documentation)
3 starsAverage (works but needs tweaking)
2 starsPoor (significant issues)
1 starBroken (doesn’t work as advertised)
0 starsLowest possible rating (also accepted by the API)
RuleDetail
Which imagesOnly an image in your library: rating any other answers 404 “You do not have access to this image”. Import a free image first
ValueA whole number from 0 to 5; a fraction answers 422 REQUEST_SCHEMA_INVALID
Rating againReplaces your earlier rating of that image

An import answers:

{
"statusCode": 200,
"message": "Free computer image imported successfully",
"data": {}
}

The image is now in your library (GET /api/v1/images/user). A purchase answers with the price paid and, when your credentials may read the wallet balance, your remaining balance:

{
"statusCode": 200,
"message": "Computer image purchased successfully",
"data": {
"price_paid": 15,
"remaining_balance": 485
}
}
What works inside a computer
Docker daemonRuns inside the computer
Docker Compose and Docker BuildxIncluded in the dev kit
Docker imagesBuilding and running them
Docker networking and volumesWork as usual
Resource limits--memory and --cpus, and --privileged Docker containers
Multi-container applicationsVia Docker Compose

Use debian/13 as the base image when running Docker. In the Hub, open Computers, choose Get a computer, select your project and server, and open More options. Under Software, turn on Hoody kit and Developer kit, then choose Get computer. For the HTTP request below, replace {id} with your project ID and {server_id} with your server ID (create a computer).

POST /api/v1/projects/{id}/computers
{
"name": "docker-host",
"server_id": "{server_id}",
"computer_image": "debian/13",
"hoody_kit": true
}

Why Debian: it is stable enough for long-running Docker daemons, its Docker packages are well supported, it has minimal conflicts with container runtimes, and it is well tested in production environments.

With the dev kit enabled, Docker is installed and starts when a Docker command first connects to it. After restarting the computer, run docker ps to activate Docker; Docker containers using --restart=always also wait for this activation. You can use it for single containers, Docker Compose stacks, or a full orchestration platform. The dev kit is on by default when hoody_kit is true, unless you set dev_kit: false. The nesting is Hoody computer, then Docker daemon, then your application containers: an HTTP-accessible Linux machine that can itself run containerized applications.

Kind, Minikube and K3s each run a full Kubernetes cluster inside a Hoody computer. Each one needs a short setup, shown below. Start from a computer created with the dev kit enabled so Docker is already there for Kind and Minikube.

Kind runs each Kubernetes node as a Docker container, so it uses the Docker that the dev kit installs. It needs no extra settings.

  1. Install the kind release binary from Kind’s Quick Start, and kubectl from the Kubernetes install guide.
  2. Create the cluster and check it:
Terminal window
kind create cluster
kubectl get nodes

The node reports Ready after about 20 seconds. kind delete cluster removes it.

K3s needs one kubelet setting before you install it. Without it, the K3s service restarts in a loop and the log shows open /dev/kmsg: no such file or directory.

Terminal window
sudo mkdir -p /etc/rancher/k3s
sudo tee /etc/rancher/k3s/config.yaml >/dev/null <<'EOF'
kubelet-arg:
- "feature-gates=KubeletInUserNamespace=true"
EOF
curl -sfL https://get.k3s.io | sudo sh -
sudo k3s kubectl get nodes

The node reports Ready after about 10 seconds. A K3s server uses about 600 MB of memory, so give the computer enough for it plus your workloads. To remove K3s, run /usr/local/bin/k3s-uninstall.sh.

Minikube runs its node as a Docker container, like Kind. Install minikube from Minikube’s Get Started page and kubectl as for Kind, then start it with the Docker driver and the Docker container runtime:

Terminal window
minikube start --driver=docker --container-runtime=docker
kubectl get nodes

Pass --container-runtime=docker every time: with Minikube’s default runtime, containerd, no control-plane pod starts. Right after minikube start, give the cluster a few seconds before your first kubectl run: until the default service account exists, it fails with serviceaccount "default" not found.

Install Podman from Debian:

Terminal window
sudo apt-get install -y podman slirp4netns

As root, Podman works as installed. To run Podman as the default user (rootless), use the slirp4netns network, because the default rootless network does not start in a Hoody computer. Make it the default once:

Terminal window
mkdir -p ~/.config/containers
cat > ~/.config/containers/containers.conf <<'EOF'
[network]
default_rootless_network_cmd = "slirp4netns"
EOF
podman run --rm docker.io/library/alpine wget -qO- https://example.com

Or pass --network=slirp4netns on each podman run.

WorkloadImageThen
Web application stack (Node.js)debian/13Install Node.js over the terminal URL or SSH
Database computer (PostgreSQL)debian/13Install PostgreSQL with apt
The same application on AMD64 servers (Intel and AMD) and ARM64 servers (Graviton, Apple Silicon)debian/13 on bothNothing: the architecture follows the target server automatically
PracticeWhy
Name the exact imageName debian/13 rather than a bare debian, so a script keeps working if the default changes
Use debian/13 for Docker hostsWhen running Docker or Kubernetes inside a computer, start with debian/13. Its Docker packages are well supported and it conflicts less with container runtimes
Importing does not prepare a createA create accepts only debian/13, whatever your library holds, so importing an image first does not speed up later creations
QuestionAnswer
What happens if I import an image I already have?The API returns 400 Bad Request with the message You already have this image. Re-importing is not idempotent, and you do not need to: skip the import step for an image you own. Importing a paid image answers 404 Free computer image not found: buy it instead
Do purchased images work on all my servers?A purchase is tied to your account, not to one server, so it stays in your library on all of them. Computer creation still accepts only debian/13
Can I roll back to a previous image version?No. The platform carries one release, debian/13. To return to a known state, snapshot a working computer and copy from that snapshot
How do I know which image to use for my application?Use debian/13. Check that the packages your application needs are in Debian’s repository
Do I need to match the image’s architecture to the server?No. A create names only debian/13, and the computer gets its server’s architecture automatically. The architecture filter (GET /api/v1/images/public?architecture=amd64, or ?architecture=arm64) only narrows the marketplace list
Do images affect computer pricing?Computers are not charged separately. A premium marketplace image has its listed purchase price
SymptomCauseWhat to do
Computer creation fails with “image not found”, or returns 400 Invalid base imageNot a lookup failure: the name is not a base image enabled for the realm. The error lists what is: debian/13Name the canonical debian/13; debian:13, debian-13, debian 13 and debian also resolve to it. Importing the image or finding it in your library does not help, because a create accepts only base images
Image purchase fails with insufficient funds: 400 “Insufficient balance”The wallet balance is below the priceCheck the wallet balance, add funds if needed, then purchase (below)
Image purchase answers 403An auth token without the financial.wallet.transfer permission (“This token does not have permission to spend wallet funds”), or a caller that is neither the signed-in account owner nor an auth tokenBuy signed in as the account owner, or with an auth token that has financial.wallet.transfer
Computer takes longer than expected to createCreation time varies; some creates take a few minutesNothing: this is normal

Check your General Balance with hoody wallet balances general get (GET /api/v1/wallet/balances lists every balance). If it is short, add funds. Then buy with hoody images buy IMAGE_ID (POST /api/v1/images/purchase/{image_id}); hoody images list shows the ID.

Before buying again, check your library with GET /api/v1/images/user: a purchase of an image you already own is refused with 400, the same status as an insufficient balance.