On-premise deployment

Air-gapped and on-premise live chat

Some networks can't reach the public internet - defense, government, healthcare intranets, and secure enterprises among them. ZChat is self-hosted end to end, so the server, the database, the agent console, and even the AI can run entirely inside your own network. Here's exactly what runs where, and what (if anything) ever needs to reach out.

Download Free Evaluation Self-Hosted vs Cloud

Inside your network

The ZChat server runs on your own Windows or Linux host. Agents and visitors on your network reach it directly - no round trip to a vendor cloud, and nothing that depends on an outside service to keep working.

Your database, your walls

Every conversation is stored in your own SQL Server. On an isolated network that means transcripts and visitor data physically cannot leave, because there is no path out for them to take.

AI without the cloud

This is the part most tools can't do. Run the chatbot on a local Ollama model and the AI answers from inside your network - no API keys, no prompts leaving the building, no dependency on an outside provider.

What runs where

Every core component lives on your infrastructure. Outbound internet is only ever for the optional pieces.

Component	Runs on	Needs outbound internet?
Chat server	Your Windows or Linux host	No
Database	Your SQL Server	No
Agent console	Browser on your network	No
Chat widget	Served from your ZChat server	No
AI chatbot (local)	A local Ollama model on your server	No
AI chatbot (hosted, optional)	OpenAI or Anthropic with your keys	Yes - only if you choose it
External webhooks (optional)	Slack, Teams, Discord, or your endpoint	Yes - only if you enable them

Honest about the edges

Fully air-gapped is a choice you can make

The core chat never needs the internet. The only components that reach out are ones you opt into: a hosted AI model, webhooks to an external service, or emailing transcripts through an outside SMTP relay. Choose a local Ollama model, keep webhooks internal, and use an internal mail server, and nothing leaves your network at all. To see exactly what's stored and where, read where your live chat data lives.

No outbound needed

Server, database, agent console, widget, and a local Ollama chatbot all run inside your network.

Outbound only if you enable it

Hosted AI models, external webhooks, and outside SMTP are optional - each is a switch you control.

Can ZChat run without internet access?

Yes. ZChat is self-hosted, so the server, your SQL Server database, the agent console, and the chat widget all run on your own network. Visitors and agents on that network can chat with no connection to the public internet.

Does the AI chatbot need a cloud service?

Not if you choose a local model. Point ZChat's chatbot at a local Ollama model and the AI runs on your own server with nothing sent outside your network. Hosted models like OpenAI or Anthropic are optional and are the only AI path that needs outbound internet.

Can ZChat run in an air-gapped environment?

The core system can. The server, database, agent console, and widget live entirely on your infrastructure. To keep it fully air-gapped, use a local Ollama model for AI and avoid webhooks that call external services - both hosted AI and outbound webhooks are optional.

Where does my chat data go on-premise?

Nowhere it doesn't already live. Every transcript, visitor record, rating, and audit entry is written to your own SQL Server database on your infrastructure. There is no vendor cloud in the path, so your data never leaves your control.

Do I need to open outbound firewall rules?

Only for the optional pieces. The core chat needs no outbound access. You would open outbound rules only if you choose a hosted AI model, send webhooks to an external service like Slack, or email transcripts through an outside SMTP relay - each of which is your choice to enable.

Owned customer support software

Deploy live chat on your own terms, not on someone else's pricing model.

ZChat gives you the installable server, web dashboard, website widget, and desktop agent tools in one self-hosted product you buy once and keep. Run it on infrastructure you trust and connect AI only if and how you want it.

Download ZChat See Product Demo

Deployment

Install on Windows or Linux, behind IIS or Nginx, in a VM, or in Docker if that fits your stack.

Commercial model

One-time purchase, perpetual license, and no monthly per-agent bill attached to growth.

AI Flexibility

Use Ollama locally or connect OpenAI and Anthropic with your own provider accounts.