CALIBER
Quickstart
Operate CALIBER

Configuration and provider setup

The operator path for configuring CALIBER: ports, database, LLM provider and gateway, object store, workflow storage, and event backend.

OperatorDeveloperHow ToGA
PrerequisitesA CALIBER deployment plan or checkout
Reviewed 2026-08-10 · current main branch docs contract

Use this page when you need the operator path for configuring CALIBER: core runtime, ports, database, LLM provider choice, object store, workflow storage, and event backend.

At a glance

AreaRepresentative settingsDeep reference
App ports and local bring-upCALIBER_PORT, MLFLOW_PORT, MLFLOW_GATEWAY_PORTLocal bring-up
Metadata databaseCALIBER_DATABASE_URLPlatform
LLM provider and gatewayCALIBER_LLM_PROVIDER, CALIBER_LLM_BASE_URLGateways
Object storeCALIBER_OBJECT_STORE_*Object store
Workflow file storageCALIBER_WORKFLOW_STORAGE_*Storage and state
Event backendCALIBER_WORKFLOW_RUN_EVENT_BACKENDWorkflows architecture

1. Configure the core runtime first

Start with the basics that determine whether CALIBER can boot and persist state:

  • runtime ports
  • metadata database
  • object or workflow storage
  • event backend

Do not start with optional tuning before these are correct.

2. Configure provider access deliberately

The current repository exposes configurable LLM provider and gateway settings. For local development, the defaults are intentionally safer and smaller than a production deployment.

Use Config and environment reference when you need the representative variable names in one place.

3. Treat secret values as sources, not literals

The configuration surface includes *_SOURCE fields for secret-backed values. That matters because CALIBER resolves secret sources rather than encouraging raw credential literals throughout deployment configuration.

4. Verify after configuration

After any meaningful configuration change, verify:

  • the service still boots
  • readiness is healthy
  • the storage and provider dependencies are reachable

Use:

CALIBER : Contextual Adaptive Lifecycle for Intelligent Build, Evaluation, and Refinement — this page is generated from the authoritative Markdown sources in docs/ and the repository-level ARCHITECTURE.md.