Agent assist

AI reply suggestions, written for your agent

The hardest part of a busy shift isn't knowing the answer - it's typing it again for the ninth time while three other chats wait. ZChat reads the conversation and drafts a few replies your agent can send in a click. The agent stays in control: they choose, they edit, they send.

Download Free Self-Hosted AI Chatbot

Drafted from the real conversation

Suggestions come from what the visitor has actually said in this chat - not a canned list. Ask for them at any point and you get a short set of replies that fit where the conversation has got to.

The agent always sends

Picking a suggestion drops it into the message box - it does not send. Your agent reads it, adjusts the wording, adds the detail only a human knows, and then hits send.

Informed by your own docs

With a knowledge base in place, relevant passages from your documents are handed to the model, so drafts describe how your product works.

Suggestions or a chatbot - they solve different problems

ZChat has both. They are not the same feature and you can run either, both, or neither.

	Reply suggestions	AI chatbot
Who the customer talks to	A person, always	The AI first, a person on handoff
What the AI produces	A draft only the agent sees	Messages sent to the visitor
Best when	You have agents and want them faster	You want questions answered out of hours
Risk if the model is wrong	Low - a human reads it first	Higher - so handoff rules matter

If you want the AI answering directly, read AI chatbot with human handoff.

Assistance, not autopilot

Built to fail quietly

If the chatbot is switched off, the model is slow, or something goes wrong, the console shows nothing at all - no error, no empty box, no interruption. An agent mid-conversation should never be made to wait on a suggestion they didn't need.

Runs where you choose

A local Ollama model keeps every transcript on your own server. OpenAI and Anthropic are there if you prefer, with your own keys.

No metering

Part of the AI assistant at $19 per agent per month - no credits and no per-message billing.

Does the AI reply to customers on its own?

Not here. Reply suggestions are drafts for the agent: they appear in the agent's console, the agent picks one, edits it if needed, and sends it. Nothing reaches a customer without a person choosing to send it. If you do want the AI to answer first, that is the separate chatbot feature with handoff.

What does the AI use to write a suggestion?

The conversation so far. If you have a knowledge base set up, relevant passages from your own documents are included too, so the drafts reflect how your product actually works rather than a generic answer.

Does this send our conversations to an AI vendor?

Only if you choose a hosted model. Point ZChat at a local Ollama model and suggestions are generated on your own server, so no transcript leaves your network. OpenAI and Anthropic are available if you prefer them, using your own keys.

What happens if the AI is switched off or unavailable?

The console simply shows nothing. Suggestions are a convenience, so a disabled chatbot or a model that is slow to answer never blocks an agent or interrupts a live conversation.

Is this an extra paid add-on?

It is part of the AI assistant, which is $19 per agent per month on the agents you choose, with no per-message or per-credit metering. Live chat itself stays free. If you run a local model, the running cost is your own hardware.

Self-hosted 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, with live chat free to run. 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

Live chat is free with unlimited agents. Only the optional AI assistant is billed, per agent per month.

AI Flexibility

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