# How It Works This page explains Hydron's full deployment pipeline — from the moment you connect your code to when your application is live and serving traffic. ## The deployment pipeline Hydron's deployment process consists of six stages: **Source Connection**, **Analysis**, **Planning**, **Provisioning**, **Building**, and **Deployment**. ## Stage 1: Source connection You provide the source of your application through one of three methods: - **Git repository** — GitHub, GitLab, Bitbucket, or any public URL - **Docker image** — An existing image from any Docker registry - **Template** — A pre-configured stack from the template gallery For Git repositories, Hydron clones the code to analyze it. For Docker images, analysis is skipped and you proceed directly to infrastructure planning.  ## Stage 2: Code analysis The AI performs a deep analysis of your codebase: **Language & framework detection** - Identifies programming languages and their versions - Detects frameworks (Express, Django, Rails, Spring, Next.js, etc.) - Reads package managers (npm, pip, bundler, maven, etc.) **Service discovery** - Finds entry points and service boundaries - Detects monorepo structures - Identifies background workers, API servers, and frontends **Configuration extraction** - Extracts environment variable references - Finds existing Dockerfiles - Detects port configurations and health check endpoints **Dependency mapping** - Maps service-to-service dependencies - Identifies database and cache requirements - Detects external service integrations The analysis results in a list of **services** with their configurations, which you review and approve.  ## Stage 3: Infrastructure planning Based on the analysis, the AI generates an **infrastructure plan** that specifies: ### Server selection The AI recommends server specifications based on your application's needs: ~~~ Application Analysis Recommended Server ───────────────────── ────────────────────── Node.js API + React KS-LE-B 3 services 4 vCPU, 16GB RAM Low-medium traffic 500GB SSD Frankfurt, DE ~~~ Factors considered: - Number and type of services - Expected resource usage (CPU, memory, disk) - Geographic location preference - Budget constraints ### Docker configuration For each service, a Dockerfile is either detected or generated. The AI creates optimized multi-stage builds that minimize image size and build time. ### Networking The plan defines: - Port mappings for each service - Internal networking between services - External port exposure for public-facing services - Reverse proxy configuration ### Environment & security - Environment variable definitions - SSL certificate requirements - Firewall rules - Security best practices You review the complete plan in the **Infrastructure** panel and can request changes through the chat before approving.  ## Stage 4: Server provisioning Once you approve the plan, Hydron provisions your dedicated server: ~~~ Provisioning Steps ┌─────────────────────────────────────────────────┐ │ [====] Install operating system Done │ │ [====] Configure networking Done │ │ [====] Install Docker & tools Done │ │ [====] Set up firewall rules Done │ │ [====] Configure SSH access Done │ │ [== ] Pull base images Running │ │ [ ] Configure reverse proxy Pending │ │ [ ] Set up SSL certificates Pending │ └─────────────────────────────────────────────────┘ ~~~ This includes: 1. **OS installation** — Ubuntu with the latest security patches 2. **Docker setup** — Docker Engine and Docker Compose 3. **Security hardening** — Firewall rules, SSH key configuration 4. **Networking** — Internal DNS, port forwarding 5. **Tooling** — Monitoring agents, log collection Each step is executed via SSH and you can watch the progress in real-time through the **Runs** panel. ## Stage 5: Build With the server ready, Hydron builds your application: 1. **Clone repository** (if Git source) or **pull image** (if Docker source) 2. **Build Docker images** using the configured Dockerfiles 3. **Push images** to the server 4. **Verify builds** — Ensure images are valid and contain expected artifacts Build logs are streamed in real-time so you can monitor progress and catch any issues. ## Stage 6: Deployment Finally, your application is deployed: 1. **Start containers** in the correct order based on dependencies 2. **Configure networking** between services 3. **Set up reverse proxy** (Nginx/Traefik) for routing 4. **Provision SSL** certificates 5. **Run health checks** to verify everything is working 6. **Configure domains** to point to the new deployment  ## After deployment Once your app is live, you can: - **Monitor** health status and resource usage - **View logs** in real-time from any service - **Redeploy** with code changes - **Scale** by upgrading server specifications - **SSH in** for direct server access ## Architecture overview Hydron consists of several internal services working together: - **Frontend** (React) — The user interface you interact with - **Backend** (Express) — API server managing business logic - **Code Analyzer** — Service that scans and analyzes repositories - **Code Deployer** — Service that builds Docker images and deploys them - **DevOps Runner** — Orchestration service for server provisioning - **LLM** (Claude) — AI engine making intelligent decisions at each stage The AI orchestrates the entire process, making intelligent decisions at each stage based on your code and requirements.