Most chatbots stop at answering FAQs. WebChatAgent learns from your website and documents, then uses that knowledge in web chat, WhatsApp and phone calls. It can collect leads, book appointments, run API actions and hand a conversation to a person when needed. Setup takes about 10 minutes and requires no code. It is hosted in Germany, works in more than 100 languages and includes a complete white-label setup for agencies with their own logo, domain, customers and plans.

Hey Product Hunt 👋

I’m Dominik, the founder and developer of WebChatAgent.

I started building it because I kept running into the same problem. Many chatbot tools are fine for a short FAQ, but they fall apart as soon as a business wants to use them for real support. Agencies also rarely get enough control to offer the product under their own name.

WebChatAgent uses one knowledge base for website chat, WhatsApp and phone calls. It learns from a website and uploaded documents. From there it can answer questions, collect leads, book appointments, trigger API workflows or pass the conversation to a person.

A big part of the product is the agency setup. You can use your own logo and domain, manage customers and plans, and keep WebChatAgent completely in the background.

It is hosted in Germany, supports more than 100 languages and can be set up without code. There is also a free plan, so you can test the website chatbot without entering a card.

I’d really appreciate honest feedback. What would you try first, and what would the assistant have to do before you would trust it with customer conversations?

The WhatsApp and phone call thing is honestly impressive, most tools I have tried only do the basic website widget. Took me like 8 minutes to get it pulling answers from my docs, which kind of blew my mind.

Trained it on my agency site in under 10 minutes and it actually pulled the right answers from PDFs too, which surprised me. The white-label setup with custom domains looks like it would save a ton of setup time for client work.

The handoff to a real person when the bot hits its limit is a really thoughtful touch. A lot of tools overlook that and end up frustrating users, so it’s nice to see execution that respects the actual flow of a conversation.