When EU residency, HIPAA, SOC 2 boundaries, air-gapped networks, or internal-only deployments take hosted chat off the table, Zchat runs the same Dchat platform on your own infrastructure.
Most teams are fine on Dchat cloud. These are the four situations where self-hosted is the requirement, not a preference.
EU GDPR, UK data-protection requirements, industry-specific rules that prohibit visitor-conversation data leaving your jurisdiction or crossing specific borders. Self-hosted chat keeps everything on infrastructure you pick.
HIPAA-covered entities, banks, defense contractors, and government-adjacent teams that can't send customer conversations through a third-party vendor without a BAA or equivalent. A self-hosted chatbot stays inside your compliance boundary.
Internal help desks, intranets, and classified environments where public-internet SaaS isn't reachable. Self-hosted deployment runs entirely inside your network with a local LLM or a private OpenAI-compatible endpoint.
Teams running 100K+ conversations per month where per-seat SaaS pricing stops making sense. A self-hosted chat server pays off on compute you already own — your only variable cost is the AI provider.
Zchat ships the same codebase that runs Dchat cloud. You don't trade features for control.
AI-first replies, same-thread human handoff, knowledge base, URL import, site crawl, proactive triggers, canned responses, routing, branded widget — everything Dchat cloud ships.
SQL Server on your hardware or cloud VM. Visitor conversations, contacts, and knowledge content never leave infrastructure you control.
OpenAI, Anthropic, Azure OpenAI, or any OpenAI-compatible endpoint — including local LLMs for air-gapped networks. AI keys stay in your environment.
Bring your own certificates, run on your subdomains, integrate with existing SSO — the chat widget loads from your origin, not a vendor CDN.
Standard ASP.NET Core 10 on IIS, Nginx, or Kestrel. Production-ready PowerShell scripts for Windows Server. Docker images available.
Perpetual license with optional support and update subscription. No per-seat fees, no per-conversation metering — pricing scales with servers, not traffic.
Same product underneath — different operating models. Most teams start on cloud and only move if specific constraints appear.
| Decision factor | Dchat cloud | Zchat self-hosted |
|---|---|---|
| Setup time | Minutes | Hours to days |
| Infrastructure management | Dchat runs it | Your team runs it |
| Where visitor data lives | Dchat-managed | Your servers only |
| Data residency control | US / cloud-managed | Any jurisdiction |
| Custom SSO / AD integration | Enterprise tier | Full access |
| Works with local / private LLM | Not yet | Yes |
| Pricing model | Per seat, flat | One-time license + optional support |
| Best for | SaaS, e-commerce, most SMB/mid-market | Regulated industries, internal tools, high volume |
Same codebase. Dchat is the fully-managed cloud service at dchat.com. Zchat at zchat.com is the self-hosted distribution — you install it on your own servers and run it yourself.
Yes. Data export from Dchat cloud into the Zchat import format keeps workspaces, knowledge base articles, agents, and contacts intact. The widget embed code changes origin but keeps the same token.
Windows Server with IIS (or Linux with Nginx/Kestrel), SQL Server 2019+, .NET 10 ASP.NET Core runtime, TLS certificates for your subdomains, and an AI provider endpoint. A single mid-tier VM handles most small-team workloads.
Yes. Zchat's AI adapter accepts any OpenAI-compatible endpoint, so a local vLLM, Ollama, or llama.cpp server works out of the box. That's how air-gapped deployments run.
Perpetual license per server, with an optional annual subscription for updates and support. No per-seat or per-conversation fees — pricing scales with infrastructure, not traffic. Contact the team for a quote.
Yes, with an active update subscription. Zchat releases track Dchat cloud; you pull the new version and run the deploy script when you're ready. No forced upgrades.
Zchat ships the same codebase Dchat cloud runs — on your infrastructure, in your jurisdiction, with your AI provider. Visit zchat.com or email the team for a deployment quote.