Skip to content

Caber Demo CLI Reference (demo.py) ​

The unified runner demo.py in CaberDemos orchestrates Docker Compose stacks, dataset seeding, automated verification, and cloud deployments across all four active demo scenarios (agentic-tools, hipaa-ner, agentgov-mimic, graphrag).

Caber Demos Pipeline
Figure 1: Unified Demo Runner Lifecycle from Seeding to Verification
## 1. Interactive & Human Workflows

These commands are designed for interactive evaluation and engineering walk throughs:

demo.py list ​

Lists all available demo scenarios defined in catalog.yaml along with their status, ports, and default datasets.

demo.py info <demo> ​

Displays detailed configuration for a scenario, including Docker services, port mappings, and environment variables.

demo.py check [demo] ​

Verifies prerequisite tooling on the host machine, including Docker Engine, Docker Compose v2, Python version, and network port availability.

demo.py up <demo> ​

Launches the demo environment and seeds initial data.

FlagTypeDefaultDescription
--no-visibleFlagFalseRun browser interactions in headless mode
--slow-moInteger300Delay browser actions by N milliseconds for demonstration visibility
--no-keep-openFlagFalseClose browser window automatically after seeding completes
--no-seedFlagFalseSkip initial dataset seeding
--dashboardFlagFalseLaunch CaberUserConsole dashboard alongside the demo
--datasetStringNoneNamed dataset from data/ to mount at /demo-data

demo.py down <demo> ​

Tears down Docker containers, volumes, and networks associated with the specified scenario.

2. CI & Automated Verification Workflows ​

demo.py ci <demo> ​

Executes full lifecycle testing: launch, seed, test assertions, and teardown.

FlagTypeDefaultDescription
--output-dirPathNoneDirectory to store test artifacts and logs
--dashboardFlagFalseLaunch CaberUserConsole for end to end audit verification
--generateFlagFalseGenerate fresh ground truth data via CaberDataFactory before launch
--verdictFlagFalseEnable deployed NER verdict pass and enforce GLiNER validation gates
--warn-on-auth-gapFlagFalseDowngrade auth correctness errors to visible warnings

demo.py test <demo> ​

Executes smoke tests and policy enforcement checks against an already running demo stack.

3. Multimedia, Build & Cloud Workflows ​

demo.py video <demo> ​

Records an automated video walkthrough with synchronized narration:

  • --tts: Text to speech engine (openai or macOS say)
  • --resolution: Output resolution (4k or 1080p)

demo.py build [demo] ​

Builds demo application containers or core platform services (--platform).

demo.py deploy <demo> & undeploy <demo> ​

Provisions demo infrastructure into AWS (--target aws) with optional automatic teardown (--ttl 4h).