diff --git a/.env b/.env
index 84087c6..e539d14 100644
--- a/.env
+++ b/.env
@@ -1,9 +1,15 @@
 POSTGRES_PASSWORD=braindb2026
+OLLAMA_URL=http://192.168.1.217:11434
 EMBEDDING_MODEL=nomic-embed-text
 CHAT_MODEL=qwen3:8b
 LOG_LEVEL=INFO
 SEARCH_TOP_K=10
 SEARCH_THRESHOLD=0.65
-CORS_ORIGINS=http://localhost:3000,http://192.168.1.16:3001
+CORS_ORIGINS=*
 CHUNK_SIZE=700
 CHUNK_OVERLAP=70
+VAULT_PATH=./vault
+PUBLIC_API_URL=http://192.168.1.16:8001
+API_PORT=8001
+UI_PORT=3001
+POSTGRES_PORT=5433
diff --git a/docker-compose.yml b/docker-compose.yml
index 8415458..9afdbf2 100644
--- a/docker-compose.yml
+++ b/docker-compose.yml
@@ -15,7 +15,7 @@ services:
       - postgres_data:/var/lib/postgresql/data
       - ./infra/database/schema.sql:/docker-entrypoint-initdb.d/01_schema.sql:ro
     ports:
-      - "${POSTGRES_PORT:-5432}:5432"
+      - "${POSTGRES_PORT:-5433}:5432"
     networks:
       - brain-net
     healthcheck:
@@ -41,53 +41,6 @@ services:
       timeout: 5s
       retries: 5
 
-  # ---------------------------------------------------------------------------
-  # Ollama (local LLM inference)
-  # ---------------------------------------------------------------------------
-  ollama:
-    image: ollama/ollama:latest
-    container_name: second-brain-ollama
-    restart: unless-stopped
-    volumes:
-      - ollama_data:/root/.ollama
-    ports:
-      - "${OLLAMA_PORT:-11434}:11434"
-    networks:
-      - brain-net
-    deploy:
-      resources:
-        reservations:
-          devices:
-            - driver: nvidia
-              count: all
-              capabilities: [gpu]
-    healthcheck:
-      test: ["CMD", "curl", "-f", "http://localhost:11434/api/tags"]
-      interval: 30s
-      timeout: 10s
-      retries: 5
-      start_period: 60s
-
-  # ---------------------------------------------------------------------------
-  # Ollama model bootstrap (pulls required models on first start)
-  # ---------------------------------------------------------------------------
-  ollama-bootstrap:
-    image: ollama/ollama:latest
-    container_name: second-brain-ollama-bootstrap
-    depends_on:
-      ollama:
-        condition: service_healthy
-    volumes:
-      - ollama_data:/root/.ollama
-    networks:
-      - brain-net
-    entrypoint: ["/bin/sh", "-c"]
-    command:
-      - |
-        OLLAMA_HOST=ollama:11434 ollama pull ${EMBEDDING_MODEL:-nomic-embed-text}
-        OLLAMA_HOST=ollama:11434 ollama pull ${CHAT_MODEL:-mistral}
-    restart: "no"
-
   # ---------------------------------------------------------------------------
   # RAG API (FastAPI)
   # ---------------------------------------------------------------------------
@@ -97,18 +50,20 @@ services:
       dockerfile: Dockerfile
     container_name: second-brain-rag-api
     restart: unless-stopped
-    env_file:
-      - .env
     environment:
       DATABASE_URL: postgresql://brain:${POSTGRES_PASSWORD:-brain}@postgres:5432/second_brain
-      OLLAMA_URL: http://ollama:11434
+      OLLAMA_URL: ${OLLAMA_URL:-http://192.168.1.217:11434}
+      EMBEDDING_MODEL: ${EMBEDDING_MODEL:-nomic-embed-text}
+      CHAT_MODEL: ${CHAT_MODEL:-qwen3:8b}
+      LOG_LEVEL: ${LOG_LEVEL:-INFO}
+      SEARCH_TOP_K: ${SEARCH_TOP_K:-10}
+      SEARCH_THRESHOLD: ${SEARCH_THRESHOLD:-0.65}
+      CORS_ORIGINS: ${CORS_ORIGINS:-*}
     depends_on:
       postgres:
         condition: service_healthy
-      ollama:
-        condition: service_healthy
     ports:
-      - "${API_PORT:-8000}:8000"
+      - "${API_PORT:-8001}:8000"
     networks:
       - brain-net
     healthcheck:
@@ -127,19 +82,18 @@ services:
       dockerfile: Dockerfile
     container_name: second-brain-ingestion
     restart: unless-stopped
-    env_file:
-      - .env
     environment:
       DATABASE_URL: postgresql://brain:${POSTGRES_PASSWORD:-brain}@postgres:5432/second_brain
-      OLLAMA_URL: http://ollama:11434
+      OLLAMA_URL: ${OLLAMA_URL:-http://192.168.1.217:11434}
       VAULT_PATH: /vault
+      EMBEDDING_MODEL: ${EMBEDDING_MODEL:-nomic-embed-text}
+      CHUNK_SIZE: ${CHUNK_SIZE:-700}
+      CHUNK_OVERLAP: ${CHUNK_OVERLAP:-70}
     volumes:
-      - ./vault:/vault:ro
+      - ${VAULT_PATH:-./vault}:/vault:ro
     depends_on:
       postgres:
         condition: service_healthy
-      ollama:
-        condition: service_healthy
     networks:
       - brain-net
 
@@ -152,19 +106,16 @@ services:
       dockerfile: Dockerfile
     container_name: second-brain-agents
     restart: unless-stopped
-    env_file:
-      - .env
     environment:
       DATABASE_URL: postgresql://brain:${POSTGRES_PASSWORD:-brain}@postgres:5432/second_brain
-      OLLAMA_URL: http://ollama:11434
+      OLLAMA_URL: ${OLLAMA_URL:-http://192.168.1.217:11434}
       VAULT_PATH: /vault
+      CHAT_MODEL: ${CHAT_MODEL:-qwen3:8b}
     volumes:
-      - ./vault:/vault:ro
+      - ${VAULT_PATH:-./vault}:/vault:ro
     depends_on:
       postgres:
         condition: service_healthy
-      rag-api:
-        condition: service_healthy
     networks:
       - brain-net
 
@@ -178,19 +129,17 @@ services:
     container_name: second-brain-ui
     restart: unless-stopped
     environment:
-      NEXT_PUBLIC_API_URL: http://localhost:${API_PORT:-8000}
+      NEXT_PUBLIC_API_URL: ${PUBLIC_API_URL:-http://localhost:8001}
     depends_on:
-      rag-api:
-        condition: service_healthy
+      - rag-api
     ports:
-      - "${UI_PORT:-3000}:3000"
+      - "${UI_PORT:-3001}:3000"
     networks:
       - brain-net
 
 volumes:
   postgres_data:
   redis_data:
-  ollama_data:
 
 networks:
   brain-net: