Environment Variables

All application config is loaded from AWS SSM Parameter Store at startup (see common/src/utils/SSMParams.js). For local development, set LOCAL=true in .envrc and the loader falls back to process.env.

API Parameters

SSM path prefix: {SSM_PARAMETER_PATH}/api/ (e.g. /dev/dgi-ai-widgets/api/)

Name	SSM Type	Description	Default
API_HOST	String	Bind address	0.0.0.0
API_PORT	String	Listen port	3000
LOG_LEVEL	String	Pino log level	info
MONGO_CONNECTION_STRING	SecureString	MongoDB connection URI	—
MONGO_MAX_POOL_SIZE	String	Mongoose connection pool size	5
DATABASE_NAME	String	Primary DB name (canvases, widgets, conversations)	dgi-ai-charts
AUTH_DATABASE_NAME	String	Auth DB name (read-only mirror from dtcea)	dtcea-mumbai-demo
AUTH_DB_MAX_POOL_SIZE	String	Auth DB connection pool size	5
CERBOS_URL	String	Cerbos gRPC endpoint	localhost:3593
albIpCidr	String	ALB IP CIDR for trusted proxy	(empty)
CLAUDE_API_KEY	SecureString	Anthropic API key for AI chart generation	—
DEFAULT_LLM	String	Primary model ID for AI chart/table generation	—
FALLBACK_LLM	String	Fallback model ID — used if primary fails or returns no tool use	—
DATA_DIR	String	Path to DuckDB data directory	/data
CANVAS_LIMIT_PER_USER	String	Max canvases a non-admin user can create	15
GCS_BUCKET	String	GCS bucket name for DuckDB data file	grid-intelligence-demo
GCS_OBJECT_KEY	String	GCS object path for DuckDB data file	duckdb/data.duckdb

Local-Only Variables

These are set in .envrc but are not stored in SSM — they control local dev behavior or AWS credential bootstrapping:

Name	Description
NODE_ENV	Node environment (development)
LOCAL	Set to true to bypass SSM and use process.env
TEST_MONGO_URI	MongoDB URI for test database
SSM_PARAMETER_PATH	SSM path prefix (e.g. /dev/dgi-ai-widgets)
ROLE_ARN_AWS	AWS IAM role ARN for web identity federation
AUDIENCE_SERVICE_ACCOUNT	GCP audience for ID token
TARGET_SERVICE_ACCOUNT	GCP service account to impersonate locally
CLOUDSDK_ACTIVE_CONFIG_NAME	gcloud config profile name
FETCH_CONFIG_FROM_S3	Script workspace: fetch config from S3 or local file

Adding a New Variable

When introducing a new environment variable, complete all of these steps:

1. SSM Parameter Store — add the put_string or put_secure call to api/ssm/setup-dev.sh (and equivalent scripts for other environments when they exist)
2. .envrc and .envrc.example — add the variable with a dev default (.envrc) and a blank placeholder (.envrc.example)
3. api/src/utils/envVars.js — export the variable from ssmParams
4. This document — add a row to the table above with name, SSM type, description, and default