Deploy and Host OpenClaw on Railway

Self-host OpenClaw — the open-source personal AI agent (formerly Clawdbot) — in one click, with a token-protected gateway + Control UI, persistent memory, and a bundled private SearXNG web-search backend. Onboard entirely in your browser: pick a model provider, name your agent, connect Telegram or WhatsApp.

About Hosting OpenClaw

OpenClaw runs as a gateway service that hosts your agent, its Control UI, and its channel connections. This template pins the current stable release (2026.7.1 — which post-dates the early-2026 security advisories) and ships with authentication always on: a 48-character gateway token is generated per deploy, and the gateway refuses unauthenticated access by design. Agent state — config, provider keys, memory, channel sessions, workspace files — lives on a persistent volume at /data, so your agent survives redeploys and upgrades. A second, private-only SearXNG service gives your agent real web search with no API key and no third-party search provider seeing your queries.

Common Use Cases

A personal AI assistant on Telegram/WhatsApp that remembers context and runs 24/7
An always-on agent that researches, summarizes, and executes recurring tasks
A private alternative to hosted agent platforms — your keys, your data, your infra

Dependencies for OpenClaw Hosting

SearXNG (included, private-only metasearch backend for the agent's web_search tool)
A model provider API key: Anthropic, OpenAI, OpenRouter, or Gemini (enter during onboarding or as a variable)

Deployment Dependencies

Implementation Details

Security posture this template ships with:

OPENCLAW_GATEWAY_TOKEN auto-generated (48 chars) — the Control UI asks for it on first open; copy it from the service's Variables tab
Pinned image ghcr.io/openclaw/openclaw:2026.7.1 — post-dates the patched CVEs from early 2026; upgrades are explicit, never silent
SearXNG is private-only (Railway private networking, IPv6) with its JSON API enabled for the agent — it is never exposed to the internet
Provider API keys are optional user-supplied variables, encrypted at rest by Railway — nothing is baked in
State on a persistent volume at /data; healthcheck on /healthz

Getting started: deploy, open https:///openclaw, paste the gateway token, and follow the onboarding wizard. Recommended resources: 2 GB memory.

Why Deploy OpenClaw 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 OpenClaw 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.