#Quick Start

This chapter walks you through the fastest path to deploy Data Agent, connect data, and validate the experience. Follow the steps to spin up an agent that can answer baseline business questions.

#1. Environment Setup

• Install Docker and Docker Compose, and make sure they run from your terminal (download: https://www.docker.com/).

#2. Fetch & Launch Services

Create docker-compose.yml in any directory (a dedicated folder makes later maintenance easier). Choose the “Global” or “Mainland China” version based on your network.

version: "3"

services:
  app:
    image: chatbi/yiask:latest
    container_name: yiask
    environment:
      - PROJECT=test
      # Configurations for the Postgres metadata database
      - DB_HOST=postgres
      - DB_DATABASE=mydatabase
      - DB_USER=postgres
      - DB_PASSWORD=postgres

      # MongoDB settings (will be deprecated in future releases)
      - MONGO_URL=mongodb://root:rootPassXXX@mongo:27017
    volumes:
      - ./storage:/app/nocobase/storage
    links:
      - mongo
      - postgres
    ports:
      - "3052:80"
    init: true
    extra_hosts:
      - "host.docker.internal:host-gateway"
    
  postgres:
    container_name: yiask_postgres
    image: pgvector/pgvector:pg17-trixie
    restart: always
    command: postgres -c wal_level=logical
    environment:
      POSTGRES_USER: postgres
      POSTGRES_PASSWORD: postgres
      POSTGRES_DB: mydatabase
    volumes:
      - ./storage/db/postgres:/var/lib/postgresql/data
      
  mongo:
    container_name: yiask_mongo
    restart: always
    image: docker.1ms.run/mongo:8.2
    environment:
      - MONGO_INITDB_ROOT_USERNAME=root
      - MONGO_INITDB_ROOT_PASSWORD=rootPassXXX
    volumes:
      - ./storage/db/mongo:/data/db
    # Uncomment to expose the port if needed
    # ports:
    #   - "27017:27017"

version: "3"

services:
  app:
    image: docker.1ms.run/chatbi/yiask:latest
    container_name: yiask
    environment:
      - PROJECT=test
      # Configurations for the Postgres metadata database
      - DB_HOST=postgres
      - DB_DATABASE=mydatabase
      - DB_USER=postgres
      - DB_PASSWORD=postgres

      # MongoDB settings (will be deprecated in future releases)
      - MONGO_URL=mongodb://root:rootPassXXX@mongo:27017
    volumes:
      - ./storage:/app/nocobase/storage
    links:
      - mongo
      - postgres
    ports:
      - "3052:80"
    init: true
    extra_hosts:
      - "host.docker.internal:host-gateway"
    
  postgres:
    container_name: yiask_postgres
    image: docker.1ms.run/pgvector/pgvector:pg17-trixie
    restart: always
    command: postgres -c wal_level=logical
    environment:
      POSTGRES_USER: postgres
      POSTGRES_PASSWORD: postgres
      POSTGRES_DB: mydatabase
    volumes:
      - ./storage/db/postgres:/var/lib/postgresql/data
      
  mongo:
    container_name: yiask_mongo
    restart: always
    image: docker.1ms.run/mongo:8.2
    environment:
      - MONGO_INITDB_ROOT_USERNAME=root
      - MONGO_INITDB_ROOT_PASSWORD=rootPassXXX
    volumes:
      - ./storage/db/mongo:/data/db
    # Uncomment to expose the port if needed
    # ports:
    #   - "27017:27017"

Run docker compose up -d to start all dependencies, and docker compose logs -f to watch them. After the containers are up, open http://<host-ip>:3052 (or http://localhost:3052 locally) to reach the console.

Default username/password: superadmin / admin123

#3. Connect Data Sources

Use the console navigation bar to open System Setup, choose Data Management → Data Sources on the left sidebar, and click Add to create a new connection. Configure DSNs, credentials, and timeout policies for your databases or APIs, then run the built-in connectivity test and save the accessible tables.

Sample connection for demo purposes:

#4. Build the Semantic Model

Open Business Modeling from the top navigation, click the Sage avatar (the AI assistant) on the right side of the toolbar, and enter “Help me build the model.” Sage will walk you through the modeling workflow step by step.

1. Define subject domains (Sales, Operations, Finance, etc.) and clarify the dimensions (time, region, channel) and metrics they need.
2. Use the metric wizard to import SQL snippets or drag-and-drop builders to generate formulas.
3. Add business terms, aliases, and sample questions so the LLM can map user intent more accurately.

#5. Run a Q&A Check

Open Intelligent Q&A from the top navigation, pick the default agent “Alisa,” and ask business questions such as “What is the sales amount?” Use the live responses to confirm accuracy; if anything looks off, return to the semantic modeling or data configuration steps to adjust.

Once you complete these steps, you will have a Data Agent that can serve business teams. Continue with the “Guide” chapter to dive deeper into data modeling, permissions, AI employees, and operations practices.