Docker Deployment

The recommended way to deploy Legible for production use.

Prerequisites

• Docker 20.10+ with Docker Compose v2
• 8 GB RAM minimum (16 GB recommended)
• 20 GB disk space

Quick Deploy

git clone https://github.com/kubeworkz/legible.git
cd legible
cp docker/.env.example docker/.env
cp docker/config.example.yaml docker/config.yaml

# Edit .env with your API keys and settings
vim docker/.env

# Build and start
./start.sh

Services

The Docker Compose stack includes:

Service	Image	Memory Limit	Purpose
wren-ui	wren-ui:local	4 GB	Next.js UI
wren-ai-service	wren-ai-service:local	4 GB	AI pipeline
ibis-server	wren-engine-ibis:local	—	Query engine + MCP
wren-engine	wren-engine:local	—	Legacy SQL engine
qdrant	qdrant/qdrant:v1.11.0	2 GB	Vector store
text-embeddings-inference	michaelf34/infinity:latest	2 GB	Embedding model
bootstrap	wren-bootstrap:local	—	One-shot init

Volumes

• data — Shared volume for MDL files, SQLite database, Qdrant storage
• tei-models — Cached embedding model weights

Exposed Ports

HOST_PORT=3000              # Wren UI
AI_SERVICE_FORWARD_PORT=5555 # AI Service API
MCP_SERVER_PORT=9000         # MCP endpoint
MCP_WEB_UI_PORT=9001         # MCP config UI

Rebuilding Individual Services

To rebuild and restart a single service:

cd docker

# Rebuild wren-ui
DOCKER_BUILDKIT=1 docker build -t wren-ui:local ../wren-ui
docker compose up -d --no-deps --force-recreate wren-ui

# Rebuild AI service
docker build -t wren-ai-service:local -f ../wren-ai-service/docker/Dockerfile ../wren-ai-service
docker compose up -d --no-deps --force-recreate wren-ai-service

Resource Tuning

For smaller machines, reduce memory limits in docker-compose.yaml:

deploy:
  resources:
    limits:
      memory: 2g  # Reduce from 4g

The minimum viable configuration needs about 6 GB RAM total.

Logs

cd docker

# All services
docker compose logs -f

# Specific service
docker compose logs -f wren-ui
docker compose logs -f wren-ai-service

Backup

The SQLite database and MDL files are stored in the data Docker volume:

# Backup
docker run --rm -v legible_data:/data -v $(pwd):/backup busybox tar czf /backup/legible-backup.tar.gz /data

# Restore
docker run --rm -v legible_data:/data -v $(pwd):/backup busybox tar xzf /backup/legible-backup.tar.gz -C /