Skip to main content

Quick Start

Let's discover LAYRA in under 10 minutes - the visual-native AI agent engine that sees, understands, and acts.


Getting Started

Deploy LAYRA with Docker to unlock its full visual-first RAG and Workflow capabilities:

What You'll Need

  1. Docker and Docker Compose
    (Installation Guide)
  2. NVIDIA GPU with drivers and NVIDIA Container Toolkit
    (Setup Instructions)
  3. 15GB+ free disk space (for AI models)

Deployment Steps

1. Clone Repository & Configure
git clone https://github.com/liweiphys/layra.git
cd layra
vim .env # Edit key parameters (SERVER_IP, MODEL_BASE_URL)
2. Launch Services (First run downloads ~15GB models, be patient)
docker compose up -d --build
CAUTION

If you encounter issues with docker compose, try using docker-compose (with the dash) instead. Also, ensure that you're using Docker Compose v2, as older versions may not support all features. You can check your version with docker compose version or docker-compose version.

3. Monitor Model Weight Initialization
docker compose logs -f model-weights-init

Be patient - Initial model download takes 20-40 mins depending on network speed. Grab a coffee!


Verify Installation

  1. Check running services:

After running docker compose up -d --build, check container status with:

docker compose ps -a

Successful installation will show:

NAME                         IMAGE                                      STATUS                      
layra-backend-1 layra-backend Up (healthy)
layra-frontend-1 layra-frontend Up
layra-kafka-1 bitnami/kafka:3.8.0 Up (healthy)
layra-kafka-init-1 layra-kafka-init Exited (0)
layra-milvus-etcd-1 quay.io/coreos/etcd:v3.5.18 Up (healthy)
layra-milvus-minio-1 minio/minio Up (healthy)
layra-milvus-standalone-1 milvusdb/milvus:v2.5.6 Up (healthy)
layra-minio-1 minio/minio Up (healthy)
layra-model-server-1 layra-model-server Up (healthy)
layra-model-weights-init-1 layra-model-weights-init Exited (0)
layra-mongodb-1 mongo:7.0.12 Up (healthy)
layra-mysql-1 mysql:9.0.1 Up (healthy)
layra-nginx-1 nginx:alpine Up
layra-python-sandbox-1 python-sandbox Exited (0)
layra-redis-1 redis:7.2.5 Up (healthy)
Key indicators of success
  1. 10+ containers in Up (healthy) status
  2. Init containers (kafka-init, model-weights-init, python-sandbox) show Exited (0) (successful exit)
  3. No error messages in status column
  4. Nginx exposes port 80 (web UI access)

Access Your LAYRA Instance

Open your browser and visit:
http://<your-server-ip> (Default port 80)

You should see:

LAYRA Interface


Troubleshooting Tips

If services fail to start:

# Check container logs:
docker compose logs <container name>

Common fixes:

nvidia-smi  # Verify GPU detection
docker compose down && docker compose up --build # preserve data to rebuild
docker compose down -v && docker compose up --build # ⚠️ delete all data to full rebuild
Take care

-v flag: Permanently delete all databases and model weights.

Choose the operation you need:

ScenarioCommandEffect
Stop services (preserve data)docker compose stopStops containers but keeps them intact
Restart after stopdocker compose startRestarts stopped containers
Rebuild after code changesdocker compose up -d --buildRebuilds images and recreates containers
Recreate containers (preserve data)docker compose down docker compose up -dDestroys then recreates containers
Full cleanup (delete all data)docker compose down -v⚠️ Destroys containers and deletes volumes
Pro Tip

For faster subsequent launches, run docker compose stop instead of down to preserve models.


Congratulations! You're now running the world's first visual-native AI agent engine. 🚀

Next Steps

  1. Upload documents in Knowledge Base
  2. Try visual Q&A in Chat
  3. Build agents in Workflow