Deploy and Host OpenHands Coding Agent on Railway

OpenHands is an open-source, model-agnostic platform for autonomous AI software development. Unlike coding assistants that suggest snippets, OpenHands runs agents that plan, write, debug, and apply changes across your entire codebase — end to end — using any frontier LLM you choose.

About Hosting OpenHands Coding Agent

Hosting OpenHands requires a sandboxed Docker runtime environment where agents can safely execute terminal commands, edit files, install dependencies, and browse the web without touching your host system. The deployment exposes a browser-accessible web UI for interacting with your AI engineer, a persistent volume to retain workspace state across restarts, and environment variables for connecting your preferred LLM provider (Anthropic, OpenAI, Google, DeepSeek, or any OpenAI-compatible endpoint). Railway handles the container orchestration, public URL provisioning, and volume management so you can focus on running tasks rather than managing infrastructure.

Common Use Cases

• Autonomous bug fixing — assign GitHub issues or failing tests to the agent and receive production-ready code changes without manual intervention
• Codebase refactoring and upgrades — automate dependency bumps, lint fixes, legacy refactors, and vulnerability remediation across one or many repositories
• Full-stack feature development — describe a feature in plain English and let the agent scaffold, implement, and test it across your frontend and backend

Dependencies for OpenHands Coding Agent Hosting

• Docker — OpenHands requires Docker socket access to spin up the isolated agent sandbox runtime container
• LLM API Key — an API key from your chosen LLM provider (Anthropic Claude, OpenAI, Google Gemini, etc.) passed via environment variable
• Persistent Volume — a Railway volume mounted to retain workspace files, agent config, and runtime state across deployments

Deployment Dependencies

• OpenHands GitHub Repository
• OpenHands Documentation
• Supported LLM Providers
• OpenHands Docker Image

Implementation Details

Set the following environment variables after deployment:

SANDBOX_RUNTIME_CONTAINER_IMAGE=docker.all-hands.dev/all-hands-ai/runtime:0.17-nikolaik
LLM_MODEL=claude-sonnet-4-20250514   # or your preferred model
LLM_API_KEY=your_api_key_here
LOG_ALL_EVENTS=true

The agent runtime runs in a child Docker container isolated from the main service. Ensure the Railway service has Docker socket access (/var/run/docker.sock) enabled for the sandbox to start correctly.

OpenHands vs. Alternatives

OpenHands competes in the autonomous coding agent category — tools that complete entire engineering tasks end-to-end, not just suggest code inline. Here is how it stacks up against the closest alternatives:

Devin is the closest direct competitor — a fully autonomous AI engineer built by Cognition AI. It is powerful but closed source, cloud-only, and priced at ~$500/month, making it inaccessible for individual developers and small teams. OpenHands was explicitly built as the open alternative.

GitHub Copilot Coding Agent converts GitHub issues into pull requests and is deeply integrated into the GitHub ecosystem. It is the right choice if your team lives entirely inside GitHub and wants zero infrastructure overhead. It does not support self-hosting, custom models, or running arbitrary terminal commands.

Claude Code is a terminal-native agent from Anthropic with exceptional reasoning and the largest context window of any tool in this space. It is not a cloud platform — it runs on the developer's local machine. It cannot be self-hosted as a shared team service and does not support BYO models.

OpenHands is the choice when you need an autonomous coding agent you fully control: open source, self-hosted, model-agnostic, with a browser UI accessible to your whole team and no per-seat licensing costs.

Why Deploy OpenHands Coding Agent 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 OpenHands Coding Agent 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.