AnChatPreview

AnChat product

Everything you need to add AI support to your product.

One place to connect knowledge, shape your assistant, and design helpful conversations. Explore the intended workflow through detailed product previews.

Try the interactive demo Read the integration guide

01 / AI CHAT WIDGET

Useful help, right inside your website.

Give visitors a simple way to ask questions without leaving the page. The widget is designed to sit comfortably alongside your existing content and adapt to smaller screens.

  • A clear welcome message that sets expectations
  • A consistent chat experience across desktop and mobile
  • Relevant answers based on the knowledge you provide

AnChat Assistant

Example conversation

PREVIEW

Hi! How can I help you today?

Do you offer free shipping?

Yes. We offer free shipping for orders over $50.

Shipping Policy · Delivery options

Illustrative answer from sample business knowledge

Try this conversation

Powered by AnChat

02 / KNOWLEDGE BASE

Teach it what makes your business yours.

Your FAQ, catalog, policies, and how-to documents already contain valuable answers. Bring those sources into one assistant so visitors can ask for information in their own words.

  • Website pages for product and service information
  • PDFs and documents for guides and policies
  • FAQs and plain text for concise, carefully reviewed answers
Knowledge sourcesSample workspace
Search business knowledge

Product catalog

Website · 12 example pages

Ready

Shipping & returns

PDF · 4 example pages

Ready

Frequently asked questions

Text · 18 example answers

Ready

Keep source material current so your assistant has the right information to work with.

03 / AI KNOWLEDGE WORKFLOW

Relevant answers start with relevant sources.

The intended workflow retrieves useful information from your connected knowledge and uses it to compose an answer. Adding content is not the same as training a foundation model from scratch.

  • Start with a narrow set of reliable sources
  • Test common questions and questions with no known answer
  • Review and update content as your business changes
Illustrative answer flow

01 · Customer question

Do you have decaf coffee?

02 · Relevant source

Product Catalog → Colombia Decaf

03 · Grounded answer

Available in 250g and 1kg bags.

When a source does not contain the answer, a clear fallback is better than an invented policy.

04 / CUSTOMIZATION

Your assistant should feel like your team.

Set the identity and tone customers see first. A familiar name, brand color, and concise welcome message make the assistant part of your product experience.

  • Name, logo, and brand color
  • Welcome message and suggested conversation starters
  • Behavior instructions and a fallback contact path
Assistant settingsPreview

Green Coffee Assistant

Your brand, in every conversation

Assistant name

Green Coffee

Tone of voice

Warm and helpful

Brand color

#4F46E5

Welcome message

Hi! Let’s find your next favorite coffee.

Behavior: use the provided catalog. If the answer is missing, say so and point to the contact page.

05 / CONVERSATION HISTORY

Understand the question behind the question.

Review how visitors ask for help and what the assistant says in return. Conversation history is designed to help you discover confusing policies and missing documentation.

  • Read a question and its answer in context
  • Spot recurring themes that deserve better source content
  • Use real questions to improve your testing checklist
Conversation historySample data

Shipping question

Example conversation

Decaf recommendations

Example conversation

Wholesale inquiry

Example conversation

Do you offer free shipping?

Yes. We offer free shipping for orders over $50.

Shipping Policy · section 2

Topic: delivery

06 / ANALYTICS

Turn questions into a clearer picture.

See which topics create the most conversations and where answers could improve. The proposed Starter plan includes basic analytics; Pro includes advanced analytics.

  • Conversation volume and questions answered
  • Popular topics and potential knowledge gaps
  • Resolution reporting shown as an illustrative product concept
Conversation overviewIllustrative data

Conversations

124

Answered

108

Resolution

87%

MonTueWedThuFriSatSun
Most common topicShipping & delivery

Sample metrics only. Resolution is illustrated as 108 ÷ 124, rounded. These are not customer results.

07 / LEAD COLLECTION

Give a good conversation a next step.

For questions that deserve a follow-up, the assistant can be designed to collect useful contact details in context. Explain what the details are for and keep the request brief.

  • Name and email for a relevant response
  • Optional phone number when a call is useful
  • A clear explanation before requesting personal information
Conversation to inquiryConcept

Would you like the team to follow up? Share your preferred contact details.

Name

Alex Example

Email

alex@example.com

Phone (optional)

Provided only when needed

Explain why you need these details and request only the information relevant to the inquiry.

Preview only · no lead is submitted

08 / Developer tools

A small integration. A familiar workflow.

The intended integration keeps assistant configuration separate from your application. These snippets preview the planned interface; the production SDK and endpoints are not available yet.

HTML
<script
  src="https://anchat.site/widget.js"
  data-assistant="assistant_xxx"
></script>
React / TSX
import { AnChat } from "@anchat/react";

export function Support() {
  return <AnChat assistantId="assistant_xxx" />;
}
Configuration, installation, and API status →

Built to fit in

One AI assistant. Everywhere your customers are.

Start with your website. JavaScript, React, and Next.js are the intended MVP integration paths; dedicated platform integrations follow.

MVP integration

JavaScript

A lightweight snippet for websites you already own.

MVP integration

React

Place your assistant inside a React application.

MVP integration

Next.js

Designed for a client widget within your Next.js app.

Coming soon

WordPress

A dedicated plugin for a simpler installation.

Coming soon

Shopify

A dedicated integration for your storefront.

Coming soon

Mobile apps

Extend conversations beyond the browser.

Coming soon

REST API

Build a custom interface on the proposed Pro plan.

Explore the integration guidesPreview documentation

The public site is a product preview. A production widget, SDK, and API are not distributed yet.

A thoughtful launch checklist.

Get the fundamentals right before you introduce an assistant to customers.

Start with a defined role

Choose a specific job, such as answering product questions or explaining a shipping policy. A narrow scope makes answers easier to evaluate and source content easier to maintain.

Review the knowledge you connect

Remove outdated policies, conflicting instructions, and content that should not be shared publicly. Do not upload passwords, payment details, or private customer records.

Test both common and difficult questions

Try the questions your team answers often, then test ambiguity, exceptions, and information that is missing. Check whether the assistant stays within the available sources.

Give customers a way to reach a person

Account changes, disputes, and special requests may need a human. Keep your contact instructions available and review the assistant’s behavior when an answer is uncertain.

Meet your next team member

Your customers have questions. Let AI answer them.

Explore what an assistant trained on your business knowledge could do. Start with a conversation.

Build your AI assistant Read the quick start

Free interactive preview · no account or payment required