Platform Deployment

the0 platform consists of multiple services that work together to provide bot development, deployment, and execution capabilities. This section covers how to deploy the platform itself, whether for local development, self-hosted production, or cloud environments.

Architecture Overview

A complete the0 deployment includes:

Application Services:

• API Server (NestJS/TypeScript) - REST API handling authentication, bot management, and orchestration
• Frontend (Next.js/React) - Web dashboard for bot management and monitoring
• Documentation (VitePress) - Platform documentation site
• Bot Runner (Go) - Single-process service for executing realtime bots
• Bot Scheduler (Go) - Single-process service for executing scheduled bots

Infrastructure Services:

• PostgreSQL - Primary database for users, bots, and configuration
• MongoDB - Runtime state and execution tracking
• NATS with JetStream - Event streaming and inter-service communication
• MinIO - S3-compatible storage for bot code, logs, and artifacts

The runtime services use a reconciliation loop pattern: they periodically compare desired state (from MongoDB) with actual state (running containers) and make corrections. Bots run in isolated Docker containers with resource limits.

Deployment Options

Docker Compose (Recommended)

Docker Compose provides the simplest and most tested path to running the0. A single make up command starts all services with sensible defaults. This approach works well for:

• Local development
• Small teams and evaluation
• Self-hosted production on a single host

See Docker Compose Deployment for setup instructions.

Kubernetes (Multi-Node & HA)

Kubernetes mode is for deployments that need multi-node scaling or high availability. The runtime controller manages bots as individual Pods, leveraging K8s for scheduling, health checks, and automatic restarts.

See Kubernetes Deployment for setup instructions.

Root Admin

Public registration is disabled. Every deployment must configure THE0_ADMIN_EMAIL and THE0_ADMIN_PASSWORD; the API creates or syncs that root admin on startup. See Root Admin Configuration.

Upgrading from an older deployment requires an explicit migration step. See the v1.14.0 Root Admin Migration Guide.

When to Use Each Mode

Scenario	Recommended Deployment
Local development	Docker Compose
Single-host production	Docker Compose
Multi-node clusters	Kubernetes
High availability required	Kubernetes

Hardware Requirements

The platform's resource requirements depend on the number of bots and execution frequency. As a baseline:

Development (Docker Compose):

• 4 CPU cores
• 8GB RAM
• 20GB disk space

Production (Docker Compose):

• 8+ CPU cores
• 16GB+ RAM
• 100GB+ disk space

Production (Kubernetes):

• 8+ CPU cores across nodes
• 16GB+ RAM
• 100GB+ disk space for persistent volumes

Service Communication

Services communicate through multiple channels:

• HTTP/REST - External API access and frontend-to-API communication
• NATS JetStream - Event-driven communication between API and runtime services
• Direct Database - Services access PostgreSQL and MongoDB directly for their data

The API server acts as the central coordinator, receiving requests from the frontend and CLI, then publishing events to NATS for the runtime services to process.

Security Considerations

Before deploying to production:

1. Change default credentials - All services ship with development passwords that must be changed
2. Configure JWT secrets - Generate strong secrets for API authentication
3. Configure the root admin explicitly - Set THE0_ADMIN_EMAIL and provide THE0_ADMIN_PASSWORD from protected deployment configuration
4. Enable TLS - Use HTTPS for all external endpoints
5. Network isolation - Infrastructure services should not be publicly accessible
6. Resource limits - Configure appropriate limits to prevent runaway bots

See the deployment-specific pages for detailed security configuration.