Environment Variables

Complete reference for all TruSpace configuration options.

Quick Reference

Category	Variables
Core	NODE_ENV, LOG_LEVEL
Network	CORS_ORIGIN, FRONTEND_PORT, BACKEND_PORT
IPFS	IPFS_*, CLUSTER_*
AI	OLLAMA_*, DISABLE_ALL_AI_FUNCTIONALITY
Security	JWT_SECRET, CLUSTER_SECRET

---

Core Settings

NODE_ENV

Application environment mode.

Value	Description
development	Development mode with verbose logging
production	Production mode with optimized settings

NODE_ENV=production

LOG_LEVEL

Logging verbosity level.

Value	Description
debug	All messages including debug
info	Informational messages and above
warn	Warnings and errors only
error	Errors only

LOG_LEVEL=info

---

Network Settings

CORS_ORIGIN

Allowed origins for CORS requests. Set to your domain.

# Local development
CORS_ORIGIN=http://localhost:3000

# Production
CORS_ORIGIN=https://yourdomain.com

FRONTEND_PORT

Port for the frontend service.

FRONTEND_PORT=3000

BACKEND_PORT

Port for the backend API service.

BACKEND_PORT=8000

OI_CORS_ALLOW_ORIGIN

CORS origin for Open Web UI.

OI_CORS_ALLOW_ORIGIN=http://localhost:3000

---

IPFS Settings

IPFS_PROFILE

IPFS configuration profile.

Value	Description
server	Optimized for servers (default)
lowpower	Reduced resource usage for Pi/embedded

IPFS_PROFILE=server

IPFS_BOOTSTRAP_REMOVE

Remove default IPFS bootstrap nodes (for private networks).

IPFS_BOOTSTRAP_REMOVE=true

IPFS_PATH

Custom IPFS data directory.

IPFS_PATH=/custom/ipfs/path

---

IPFS Cluster Settings

CLUSTER_SECRET

Shared secret for cluster authentication. Must be the same on all connected nodes.

Security

Generate a unique secret for production:

openssl rand -hex 32

CLUSTER_SECRET=your-32-byte-hex-secret-here

CLUSTER_PEERNAME

Human-readable name for this cluster peer.

CLUSTER_PEERNAME=node-1

CLUSTER_REPLICATION_MIN

Minimum number of nodes that should pin each item.

CLUSTER_REPLICATION_MIN=2

CLUSTER_REPLICATION_MAX

Maximum number of nodes that should pin each item.

CLUSTER_REPLICATION_MAX=3

---

AI Settings

DISABLE_ALL_AI_FUNCTIONALITY

Completely disable AI features.

# Enable AI (default)
DISABLE_ALL_AI_FUNCTIONALITY=false

# Disable AI
DISABLE_ALL_AI_FUNCTIONALITY=true

OLLAMA_MODEL

Default LLM model for AI analysis.

Model	Size	Quality	Speed
tinyllama	637 MB	Basic	Fast
phi3	2.2 GB	Good	Medium
llama3.2:3b	2.0 GB	Good	Medium
llama3.2:7b	4.7 GB	Better	Slower
mistral	4.1 GB	Better	Slower

OLLAMA_MODEL=llama3.2:3b

OLLAMA_HOST

Ollama API host (if running separately).

OLLAMA_HOST=http://localhost:11434

OLLAMA_GPU

Enable GPU acceleration.

# Auto-detect (default)
OLLAMA_GPU=auto

# Force CPU
OLLAMA_GPU=cpu

---

Security Settings

JWT_SECRET

Secret key for JWT token signing.

Security

Generate a unique secret for production:

openssl rand -hex 64

JWT_SECRET=your-very-long-secret-key-here

JWT_EXPIRY

JWT token expiration time.

JWT_EXPIRY=24h

BCRYPT_ROUNDS

Number of bcrypt hashing rounds.

BCRYPT_ROUNDS=12

---

Build Settings

BUILD_OR_PULL_IMAGES

Whether to build images locally or pull from registry.

Value	Description
pull	Pull pre-built images (default, faster)
build	Build images locally

BUILD_OR_PULL_IMAGES=pull

---

Example Configurations

Local Development

.env

NODE_ENV=development
LOG_LEVEL=debug
CORS_ORIGIN=http://localhost:3000
FRONTEND_PORT=3000
BACKEND_PORT=8000
OLLAMA_MODEL=tinyllama

Production

.env

NODE_ENV=production
LOG_LEVEL=info
CORS_ORIGIN=https://truspace.example.com
FRONTEND_PORT=3000
BACKEND_PORT=8000
JWT_SECRET=<generated-secret>
CLUSTER_SECRET=<generated-secret>
OLLAMA_MODEL=llama3.2:7b
BUILD_OR_PULL_IMAGES=pull

Raspberry Pi

.env

NODE_ENV=production
LOG_LEVEL=info
IPFS_PROFILE=lowpower
DISABLE_ALL_AI_FUNCTIONALITY=true
CORS_ORIGIN=http://raspberrypi.local:3000

---

Applying Changes

After modifying .env:

# Restart services
docker compose down
docker compose up -d

# Or use start.sh
./start.sh