Deploy and Host Heym on Railway

Heym is a self-hosted runtime and platform for building, orchestrating, running, observing and evaluating agentic systems. Visually compose AI agents and deterministic workflows (or generate them with AI), connect any LLM, tools, MCP servers, company knowledge and internal APIs, then operate them with approval checkpoints, retries, traces and evals — all on your own infrastructure.

About Hosting Heym

This template deploys:

Service	Purpose
heym	Official `ghcr.io/heymrun/heym` release image: FastAPI backend + Vue frontend in one container (port 4017). Runs database migrations automatically on boot. Persistent volume at `/app/data` for uploaded files and plugins
Postgres	Workflow definitions, runs, credentials (encrypted), users

Generated per deployment: SECRET_KEY (sessions/JWT) and ENCRYPTION_KEY (encrypts stored credentials — never change it after first boot). BACKEND_WORKERS is set to 2 to keep memory reasonable; raise it for heavy workloads.

After deploy:

Open the generated domain and register your account (sign-up is open on a fresh instance).
Add your LLM provider credentials (OpenAI, Anthropic, Gemini, OpenRouter, Ollama…) in Credentials.
Build a workflow on the canvas or generate one with AI, then deploy it as an API endpoint, chat, schedule or webhook.
Optional hardening: set ALLOW_REGISTER=false on the heym service once your team has signed up, and HEYM_ADMIN_EMAILS to your admin address(es).

Common Use Cases

Production AI agents with human-in-the-loop approvals, retries and full run history
Internal automations that mix LLM steps with deterministic API/database steps
RAG assistants over company knowledge exposed as chat or API endpoints
Evaluating and comparing agent versions before rolling them out

Dependencies for Heym Hosting

PostgreSQL (included)
API key(s) for your LLM provider(s), added in the UI after deploy

Deployment Dependencies

Source: https://github.com/heymrun/heym
Website & docs: https://heym.run

Implementation Details

Features that spawn sibling Docker containers (Codex/OpenCode coding nodes and Docker-sandboxed MCP stdio/Python tools) need a Docker socket and are not available on Railway; all core workflow, agent, LLM, HTTP, MCP (remote) and knowledge features work. Heym is MIT-licensed with the Commons Clause — review the license before offering it as a paid service.

Why Deploy Heym on Railway?

Railway is a singular platform to deploy your infrastructure stack. Railway will host your infrastructure so you don't have to deal with configuration, while allowing you to vertically and horizontally scale it.

By deploying Heym on Railway, you are one step closer to supporting a complete full-stack application with minimal burden. Host your servers, databases, AI agents, and more on Railway.