Key Concepts

Understanding Hydron's core concepts will help you make the most of the platform. This page explains the fundamental building blocks and how they fit together.

Overview

Hydron is organized around five core concepts: Projects, Services, Chats, Infrastructure Plans, and Servers. Each builds on the others to create a complete deployment workflow.

Projects

A Project is the top-level container for everything related to a deployment. It groups together your services, conversations, infrastructure plans, and deployment history.

Key characteristics:

• Each project has a name and type (Web App, Microservice, Static Site, etc.)
• Projects contain one or more services (your application components)
• Projects have chats where you interact with the AI assistant
• Each project tracks its own infrastructure plans and deployment runs

Common patterns:

• One project per application — A full-stack app with frontend, backend, and database
• One project per environment — Separate projects for staging and production
• One project per client — Freelancers managing multiple client deployments

Services

A Service represents a deployable component within your project. Services are detected during code analysis or added manually.

Examples of services:

• An Express.js API server
• A React frontend application
• A PostgreSQL database
• A Redis cache
• A background worker process

Each service has:

• A name and type identifier
• A source — Git repository, Docker image, or template
• A Dockerfile — Existing or AI-generated
• Environment variables — Configuration values the service needs
• Port mappings — Which ports the service exposes
• A health check — Endpoint and interval for monitoring

Services can depend on each other. For example, your API server might depend on a database service. Hydron understands these relationships and configures networking accordingly.

Chats

The Chat is your primary interface for interacting with Hydron. Instead of filling out forms or writing configuration files, you describe what you need in natural language.

What you can do in a chat:

• Connect a repository and trigger code analysis
• Ask the AI to generate or modify infrastructure plans
• Review and approve deployment steps
• Ask questions about your infrastructure
• Request changes to server configurations
• Troubleshoot deployment issues

How chats work:

• Each chat belongs to a project
• Messages are streamed in real-time from the AI
• The AI can trigger actions like code analysis, plan generation, and deployment
• Action cards appear in the chat for reviewing and approving steps
• Multiple chats can exist per project for different topics or deployment iterations

Infrastructure Plans

An Infrastructure Plan is the AI-generated blueprint for deploying your application. It specifies everything needed to go from code to a running production environment.

A plan includes:

Plan lifecycle:

1. Generated — AI creates the plan based on code analysis
2. Reviewed — You review the plan in the Infrastructure panel
3. Modified — Ask the AI to change any aspect
4. Approved — You approve the final plan
5. Executed — Servers are provisioned and app is deployed

Plans become stale if you change services after the plan was created. Hydron will notify you and suggest regenerating the plan.

Servers

Hydron deploys your applications to dedicated servers — real machines that you own and control. These are not shared hosting or serverless functions.

Server characteristics:

• Dedicated hardware — Your own CPU, RAM, and storage
• Full SSH access — Connect directly via SSH with root access
• Docker-based — Applications run in Docker containers
• Managed provisioning — Hydron handles OS installation and configuration
• No vendor lock-in — Your server, your data, your control

Server lifecycle:

1. Selected — You choose a server specification from available options
2. Provisioned — OS is installed and Docker is configured
3. Deployed — Your application containers are running
4. Monitored — Health checks verify your app is running

Servers are managed through the Server Inventory, where you can view details, add tags, link servers to projects, and monitor their status.

Deployment Runs

A Run represents a single execution of a deployment pipeline. Each time you deploy or redeploy, a new run is created.

A run tracks:

• Provisioning steps — Server setup commands and their status
• Build steps — Docker image building progress
• Deploy steps — Container deployment and health check results
• Logs — Real-time streaming output from each step

Run statuses:

• Started — The run has been initiated
• In Progress — Steps are actively executing
• Completed — All steps finished successfully
• Failed — One or more steps encountered an error
• Paused — The run was manually paused
• Stopped — The run was manually stopped

You can pause, resume, or stop a run at any time. Failed runs show detailed error logs to help you troubleshoot.

How everything connects

Here's how a typical deployment flows through these concepts:

1. Create a PROJECT
2. Start a CHAT within the project
3. Connect your code — AI detects SERVICES
4. AI generates an INFRASTRUCTURE PLAN
5. You review and approve the plan
6. A DEPLOYMENT RUN provisions SERVERS
7. Your app is live on dedicated SERVERS

Each concept builds on the previous one, and the AI assistant guides you through each step via the chat interface.

Next steps

• Deploy from Git — Walk through a complete deployment
• Projects — Learn how to organize and manage projects
• Infrastructure Plans — Understand plan details