Packaging Commands

Package workflows for distribution and generate deployment artifacts.

kdeps bundle package

Package workflow or component into an archive for distribution.

kdeps bundle package [directory] [flags]

Output by detected manifest:

Detected file	Output format	Extension
workflow.yaml	Workflow package	.kdeps
agency.yaml	Agency package	.kagency
component.yaml	Component package	.komponent

Flags:

Flag	Description	Default
--output, -o	Output directory	. (current)
--name	Package name	From metadata (name-version)

What it packages:

• Main manifest (workflow.yaml, agency.yaml, or component.yaml)
• All resource files (resources/)
• Python requirements (requirements.txt)
• Data files and scripts
• HTML/CSS/JS assets (for components)
• Respects .kdepsignore exclusions

Examples:

kdeps bundle package my-agent/                # Creates my-agent-1.0.0.kdeps
kdeps bundle package my-agency/               # Creates my-agency-1.0.0.kagency
kdeps bundle package my-component/            # Creates greeter-1.0.0.komponent
kdeps bundle package my-agent/ --output dist/
kdeps bundle package my-agent/ --name custom-agent

---

kdeps bundle prepackage

Bundle a .kdeps or .kagency package with the kdeps runtime into a single self-contained executable per architecture. The produced binary needs no kdeps installation: running it executes the embedded package directly.

kdeps bundle prepackage <package.kdeps|package.kagency> [flags]

Flags:

Flag	Description	Default
--output, -o	Output directory	. (current)
--arch	Single target (e.g. linux-amd64)	all targets
--kdeps-version	Runtime version to embed	running binary's version
--include-models	Pre-bake the chat models' llamafiles into the executable	off

--include-models makes the binary fully offline: every literal chat.model in the package's workflows is resolved through the llamafile registry (downloaded into your local cache if missing) and embedded under a reserved .kdeps-models/ directory. At run time the embedded models become the llamafile cache, so aliases like llama3.2:1b resolve without network. Expect roughly +1.1 GB per model in the output binary.

prepackage --include-models          run the binary
        |                                  |
        v                                  v
[kdeps runtime][workflow][model]  -->  model served from the
   one executable file                 embedded .kdeps-models/

Examples:

# All architectures
kdeps bundle prepackage my-agent-1.0.0.kdeps

# One target, fully offline (model embedded)
kdeps bundle prepackage my-agent-1.0.0.kdeps --arch linux-amd64 --include-models

# Agencies work the same way
kdeps bundle prepackage my-agency-1.0.0.kagency --include-models

---

kdeps bundle build

Build Docker image from workflow.

kdeps bundle build [path] [flags]

Accepts: directory, workflow file, or .kdeps package.

Flags:

Flag	Description	Default
--gpu	GPU type: cuda, rocm, intel, vulkan (auto-selects Ubuntu base)	None (CPU, Alpine)
--tag, -t	Docker image tag	From workflow metadata
--no-cache	Build without cache	false

Examples:

kdeps bundle build examples/chatbot                            # CPU, Alpine
kdeps bundle build examples/chatbot --gpu cuda                 # NVIDIA GPU
kdeps bundle build examples/chatbot --gpu rocm                 # AMD GPU
kdeps bundle build examples/chatbot --tag my-agent:v1.0.0
kdeps bundle build myapp-1.0.0.kdeps                          # From package

---

kdeps export iso

Export a workflow as a bootable image (ISO, raw disk, or qcow2) using LinuxKit.

kdeps export iso [path] [flags]

See kdeps export iso --help for the full list of formats and flags.

---

kdeps export k8s

Generate Kubernetes Deployment and Service manifests from a workflow.

kdeps export k8s [path] [flags]

Arguments: directory, workflow file, or .kdeps package.

Flags:

Flag	Short	Description	Default
--image	-i	Container image name	{name}:{version}
--output	-o	Output file path	stdout
--replicas	-r	Number of replicas	From workflow
--network-policy		Also generate an ingress-restricting NetworkPolicy	Off

Examples:

kdeps export k8s examples/chatbot              # Print to stdout
kdeps export k8s examples/chatbot \
  --image my-registry/chatbot:v1.0.0 \
  --output k8s.yaml                            # Save to file
kdeps export k8s examples/chatbot --replicas 5 # Override replicas

Manifests are driven by agentSettings in workflow.yaml:

• replicas -- number of pod replicas
• resources -- CPU/memory limits and requests
• env -- container environment variables
• networkPolicy: true -- appends a NetworkPolicy restricting ingress to the configured ports
• portNum inside apiServer:/webServer: -- exposed ports
• installOllama: true -- adds Ollama backend port (11434)

See Kubernetes Deployment for full details.

See Also

• CLI Overview -- global flags, exit codes, env vars
• Dev Commands -- run, serve, validate, new
• Docker Deployment
• Kubernetes Deployment