Deploy guides
End-to-end deployment walkthroughs, each labelled with estimated time and hardware.
Alibaba Cloud — full stack for China compliance
Required for China compliance — run Chatwoot + Dify on Aliyun with ICP filing, RDS, SLB and monitoring.
Run Chatwoot + Dify on Cloudflare + your own VPS
Cloudflare handles DNS, WAF and Tunnel; the VPS runs the core services. From zero to live in about 90 minutes.
Docker Swarm — the lightweight K8s alternative
When 3-5 hosts need to coordinate but you don't want K8s, Swarm is the sweet spot. Init, join, rolling upgrades, monitoring.
Chatwoot + Dify on AWS ECS Fargate
Serverless containers for AI support — no EC2, pay per resource. Terraform module, auto-scaling and cost control.
One-click AI support on Coolify — your own Vercel
Coolify is an open-source PaaS that gives your VPS a Vercel-grade DX. Run Chatwoot + Dify in 30 minutes.
Global edge deploy on fly.io
fly.io runs your Chatwoot + Dify in multiple regions and routes each request to the nearest node automatically.
Hetzner — the best price/performance choice for AI support
Same specs cost 1/3 to 1/5 of AWS / AliCloud. Sizing, Ansible automation and monitoring end to end.
Production-grade Chatwoot + Dify on Kubernetes
Multi-node K8s deploy of Chatwoot + Dify with HPA, Ingress, cert-manager and PVC — end to end.
Chatwoot + Dify on Railway
Railway turns docker-compose service graphs into one-click cloud deployments — five minutes from GitHub repo to live.
Deploy AI support on Sealos — a developer-friendly K8s "desktop"
Sealos packages Kubernetes as a macOS-like desktop console. Chinese-region deploy with no YAML.
Vercel for the frontend — LobeChat public portal at the edge
Chatwoot / Dify are stateful backends and don't belong on Vercel. But stateless frontends like LobeChat one-click deploy globally with zero ops.
Deploy AI support on Zeabur — Chinese-friendly PaaS
Zeabur is a modern PaaS built by Chinese-speaking developers — true docker-compose deploys, transparent pricing.