Configuration profiles

How profile files, environment variables, model providers, and deployment modes work.

Profile files

Vitruvian setup creates:

• .env.Vitruvian (active profile pointer)
• .env.Vitruvian.<profile> (profile-specific values)

Supported profiles:

• dev
• personal
• team
• prod

Essential variables

Variable	Required	Purpose
VITRUVIAN_PROFILE	Recommended	Selects active profile file
VITRUVIAN_MODEL_PROVIDER	Yes	openai, anthropic, gemini
OPENAI_API_KEY / ANTHROPIC_API_KEY / GEMINI_API_KEY	Conditional	Provider credentials
VITRUVIAN_MODEL_NAME	No	Override provider default model
VITRUVIAN_WORKING_DIRECTORY	No	Base workspace for file operations
VITRUVIAN_MEMORY_CONNECTION_STRING	No	Durable memory and audit storage

Deployment-mode variables

Mode	Required variables
Local CLI	none beyond provider config
Discord	DISCORD_BOT_TOKEN, DISCORD_CHANNEL_ID
WebSocket	VITRUVIAN_WEBSOCKET_URL

Optional WebSocket helpers:

• VITRUVIAN_WEBSOCKET_PUBLIC_URL
• VITRUVIAN_WEBSOCKET_DOMAIN

Non-interactive profile switching

Linux / macOS

./scripts/install.sh dev
./scripts/install.sh team

Windows PowerShell

.\scripts\install.ps1 -Profile dev
.\scripts\install.ps1 -Profile team

Gotchas

• Use lowercase provider values (openai, anthropic, gemini) when following the installation flow in this documentation.
• Do not mix old variable prefixes with new ones in the same profile.
• If setup prompts for keys repeatedly, make sure your profile file is writable.

Next: read Running Vitruvian day to day.