Environment Variables

Core

Variable	Default	Description
MODELREINS_PROVIDER	—	Default provider for workers
MODELREINS_API_URL	http://localhost:7420	Coordinator URL
MODELREINS_API_KEY	—	API key for coordinator auth
MODELREINS_CONFIG_URL	—	Remote config endpoint URL

Provider API keys

Variable	Provider
ANTHROPIC_API_KEY	Claude
OPENAI_API_KEY	OpenAI
GOOGLE_AI_API_KEY	Gemini
OPENROUTER_API_KEY	OpenRouter
ONEMINAI_API_KEY	1minAI

Local provider hosts

Variable	Default	Description
MODELREINS_OLLAMA_HOST	http://localhost:11434	Ollama server URL
MODELREINS_LMSTUDIO_HOST	http://localhost:1234	LM Studio server URL

Worker

Variable	Default	Description
MODELREINS_WORKER_NAME	hostname	Worker display name
MODELREINS_WORKER_CONCURRENCY	1	Max concurrent jobs
MODELREINS_WORKER_TAGS	—	Comma-separated worker tags

Coordinator

Variable	Default	Description
MODELREINS_PORT	7420	Coordinator and dashboard port
MODELREINS_HOST	localhost	Coordinator bind address
MODELREINS_POSTGRES_URL	—	PostgreSQL connection string (required for SaaS mode)
MODELREINS_DB_BACKEND	postgres	Database backend (postgres or sqlite for local dev only)

Provider Catalog Sync

See Provider Catalog Sync.

Variable	Default	Description
MODELREINS_CATALOG_SYNC	1	Set to 0 to disable the catalog sync loop
MODELREINS_CATALOG_INTERVAL	300	Seconds between due-subscription checks
MODELREINS_CATALOG_DIR	./providers/catalog	Where provider catalog YAML configs live
MODELREINS_CATALOG_TIMEOUT	20	Per-request timeout (seconds) for a catalog fetch

MCP Channel

Variable	Default	Description
MODELREINS_MCP_TIMEOUT	30000	Timeout in ms for sync completions
MODELREINS_RECEIVE_JOBS	true	Set false for outbound-only — dispatch/status tools stay available, but this session never polls for or executes dispatched jobs.