Quick Start

Introduction

Everything you need to know about Chatsby — the platform for building AI customer support agents trained on your own data. Learn how it works, what you can build, and how to get started.

Introduction to Chatsby

Chatsby is a platform that lets businesses build, train, and deploy AI-powered customer support agents using their own data. You provide the knowledge — help articles, product docs, PDFs, website content, internal FAQs — and Chatsby turns it into a conversational AI agent that can answer customer questions accurately, 24 hours a day, in any language your customers speak.

The problem is straightforward: support teams are expensive to scale, customers expect instant answers, and most of the questions your team handles are repetitive. Chatsby solves this by letting you create an AI agent that handles the routine questions automatically, captures leads when appropriate, and escalates to a human when the conversation requires it.

No machine learning expertise required. No infrastructure to manage. You can go from raw data to a live agent embedded on your website in under ten minutes.

Platform Capabilities

Chatsby is not a generic chatbot builder. It is a complete system for deploying AI agents that are grounded in your specific business data. Here is what the platform provides:

Multi-Source Training

Your agent's quality depends entirely on the data you feed it. Chatsby supports ingesting content from multiple source types simultaneously — website URLs (with automatic crawling of subpages), PDF documents, plain text, DOCX files, and structured Q&A pairs. You can combine all of these into a single agent's knowledge base, and retrain at any time as your content changes.

Chat Widget Embedding

Deploy your agent to any website with a single line of HTML. The widget is fully customizable — colors, position, welcome messages, avatar, and branding all match your site. It works on any platform: WordPress, Shopify, Webflow, Wix, custom-built sites, or single-page applications.

REST API

For teams building custom experiences, the Chatsby API gives you full programmatic control. Send messages, manage conversations, create and update agents, and retrieve analytics data. Build your own chat interface, integrate into a mobile app, or connect to internal tools.

Analytics Dashboard

Every conversation is logged and analyzed. The dashboard shows you total conversations, resolution rates, popular questions, unanswered queries, and user satisfaction signals. Use this data to identify knowledge gaps in your training data and continuously improve your agent's accuracy.

Team Collaboration

Invite team members with role-based access. Admins manage billing and settings. Editors update training data and agent configuration. Viewers monitor analytics and conversation logs. Built for organizations where multiple people contribute to the support experience.

Integrations

Connect your agent to the tools your team already uses. Chatsby integrates with Slack (receive notifications and respond from channels), WhatsApp (deploy your agent on WhatsApp Business), Zapier (trigger workflows from conversations), HubSpot (sync leads and contacts), and Zendesk (escalate tickets automatically). See the full list in the Integrations Overview.

Lead Capture and Human Escalation

Configure your agent to collect visitor information — name, email, company — before or during a conversation. When the AI cannot resolve an issue, it hands off to a human agent with the full conversation context preserved. No information is lost in the transition.

How It Works: Architecture Overview

Understanding the data flow helps you get the most out of the platform:

  1. Sources — You provide your business content (websites, documents, text, Q&A pairs). Chatsby processes, chunks, and embeds this content into a vector store optimized for semantic search.
  2. Training — The platform indexes your content so the AI model can retrieve the most relevant passages for any given question. This is retrieval-augmented generation (RAG), not fine-tuning — your data is never used to train the underlying model.
  3. Agent — When a user asks a question, the agent searches your indexed content, retrieves the most relevant context, and generates a grounded response using a large language model. The system prompt you configure controls tone, behavior, and guardrails.
  4. Deployment — Your agent is accessible via the embeddable widget, a shareable link, or the REST API. All three methods connect to the same agent and knowledge base.
  5. Analytics — Every interaction is recorded. You can review conversations, track metrics, and identify areas where your agent needs more training data.

Getting Started in 3 Steps

Create and Train Your Agent

Sign in to your Chatsby dashboard and click New AI Agent. Give it a name that reflects its purpose (e.g., "Acme Support Bot"). Then add your training data — paste website URLs for automatic crawling, upload PDF or DOCX files, add raw text content, or define Q&A pairs for precise control over specific answers. Click Create Agent to begin training. Processing typically takes 30 seconds to a few minutes depending on the volume of content.

Test and Refine in the Playground

Once training completes, you are taken to the Playground — a live testing environment where you can interact with your agent. Ask it questions your customers would ask. Check that responses are accurate, appropriately scoped, and match the tone you want. If something is off, adjust the system prompt, add more training data, or fine-tune the temperature setting to control response creativity. Iterate until the agent performs to your standard.

Deploy to Your Website or App

Navigate to the Connect tab, set your agent to Public, and copy the embed script. Paste it into your website's HTML before the closing </body> tag. Your agent is now live — accessible to every visitor on your site. Alternatively, use the REST API to integrate the agent into a custom application, mobile app, or internal tool.

Use Cases

Chatsby is used across industries. The table below shows common deployments with realistic outcomes:

Use CaseWhat the Agent DoesTraining DataExpected Outcome
Customer SupportAnswers product questions, explains policies, troubleshoots common issuesHelp center articles, FAQ pages, product documentation40-60% reduction in support ticket volume within the first month
E-commerce SalesRecommends products, explains shipping and returns, assists with order lookupProduct catalog, shipping policies, return policy pagesHigher conversion rates, fewer abandoned carts from unanswered questions
SaaS OnboardingGuides new users through setup, explains features, links to relevant docsProduct docs, getting started guides, changelogFaster time-to-value, reduced onboarding support load
Internal HelpdeskAnswers employee questions about HR policies, IT procedures, benefitsInternal wikis, policy documents, procedure manualsFewer repetitive tickets to HR and IT teams
Lead QualificationEngages website visitors, asks qualifying questions, captures contact infoSales collateral, pricing pages, case studiesQualified leads delivered to your CRM automatically via integrations
Education & TrainingAnswers student questions about course material, provides study guidanceCourse content, textbooks (PDF), lecture notesOn-demand tutoring available 24/7 without instructor time

Quick Integration Examples

Widget Embed (HTML)

Add this script before the closing </body> tag on any page where you want the chat widget to appear:

<script
  src="https://chatsby.co/widget.js"
  data-bot-id="YOUR_BOT_ID"
  defer
></script>
Replace YOUR_BOT_ID with the agent ID from your Connect settings page. The widget loads asynchronously and will not affect your page load speed.

API Usage (JavaScript)

Send a message to your agent and receive a response programmatically:

const response = await fetch('https://api.chatsby.co/v1/agents/YOUR_BOT_ID/chat', {
  method: 'POST',
  headers: {
    'Content-Type': 'application/json',
    'Authorization': 'Bearer YOUR_API_KEY'
  },
  body: JSON.stringify({
    message: 'What is your return policy?',
    conversationId: 'optional-existing-conversation-id'
  })
});
 
const data = await response.json();
console.log(data.message);       // The agent's response text
console.log(data.conversationId); // Use this to continue the conversation
console.log(data.sources);        // Source documents the agent referenced

API Usage (cURL)

curl -X POST https://api.chatsby.co/v1/agents/YOUR_BOT_ID/chat \
  -H "Content-Type: application/json" \
  -H "Authorization: Bearer YOUR_API_KEY" \
  -d '{"message": "How do I reset my password?"}'

For full API documentation, see the Developer API Reference.

Platform Limits

Every Chatsby plan includes generous limits. Here is what to expect:

ResourceFree PlanPro PlanEnterprise
Agents15Unlimited
Training sources per agent10100Unlimited
Messages per month10010,000Custom
File upload size10 MB50 MB200 MB
Team members15Unlimited
API accessNoYesYes
IntegrationsWidget onlyAllAll + custom
Limits vary by plan. Visit the billing FAQ for current pricing and detailed plan comparisons.

Your next step depends on your role and what you need to accomplish:

If you are a business user or team lead:

  • Your First Agent — Step-by-step walkthrough to create, train, and deploy your first agent
  • Sources — Deep dive into training data types and best practices
  • Analytics — Understanding your agent's performance metrics
  • Team Management — Inviting team members and managing roles

If you are a developer:

If you need integrations:

If you have questions: