Introduction

Picture this: someone lands on your website at 11pm on a Sunday with a question about your pricing or services. Normally they'd wait until Monday. Half leave. The other half forget they were interested.

Now imagine a chatbot that answers instantly, in plain English, with exactly the right information — handles follow-ups, stays on topic, and asks for their email so you can follow up in the morning. It runs 24 hours a day without you doing anything.

That's not science fiction in 2026. It's an afternoon project.

This guide walks you through building exactly that — a working AI chatbot, powered by Claude or GPT-4, that you can embed on any website or share as a standalone link. No coding. No server setup. No technical experience required.

TL;DR: Using Botpress (free tier) and an AI API key, you can build and deploy a working chatbot in under two hours. This guide walks you through every step.

What You'll Build

By the end of this guide, you'll have a chatbot that:

Two common use cases to help you picture it: a small business owner could build a FAQ bot that answers questions about opening hours, services, and booking — saving hours of repeated email responses each week. A freelancer could build a portfolio assistant that tells visitors about their work, rates, and availability, and captures contact details from serious inquiries.

Both of those are built with exactly the same tools and approach you'll use here.

What You Need Before You Start

The requirements for this are much shorter than you might expect:

That's genuinely it. You don't need to download anything. You don't need to set up a coding environment. You don't need a paid plan to get started. Everything in this guide works on free tiers.

The one thing that makes the biggest difference isn't technical — it's clarity. The clearer you are about what your bot is for and what it should say, the better it will perform. Spend ten minutes thinking about that before you open the browser.

Choosing Your Tools: The No-Code Chatbot Landscape

Before diving into the build, it's worth knowing what your options are. Three tools lead the no-code chatbot space in 2026.

Botpress is the best overall choice for beginners. The free tier gives you five bots and 2,000 messages per month — more than enough to start. The interface uses a visual flow builder (a flowchart you drag and drop, not code you write), with native integration for both Claude and GPT-4. Embedding your chatbot on a website takes one line of code that Botpress generates for you.

Voiceflow is the better choice for more complex conversational flows — branching paths, multi-step logic, voice interfaces. More powerful, but a steeper learning curve and a more limited free tier.

Typebot is the best open-source option. You can self-host it (run it on your own server), and it produces beautiful form-style chat interfaces — great for lead capture, surveys, and quizzes.

For this guide, we're using Botpress. Widest free tier, easiest setup, best documentation for beginners.

Step 1: Set Up Your Botpress Account

Go to botpress.com and click the sign-up button. You can register with an email address or connect a Google account — either works fine. Verify your email when the confirmation arrives.

Once you're in, you'll land on the Botpress dashboard. It looks a bit like a project management tool: a sidebar on the left with your bots and settings, and a main area that shows your workspace. If this is your first time, it'll be mostly empty.

Click Create Bot — you'll see it either as a button in the centre of the screen or in the top-right area. Give your bot a specific name. Don't call it "My Bot" or "Test Bot." Name it after what it actually does: "Acme FAQ Assistant" or "Sarah's Portfolio Bot." This matters more than it sounds — a specific name helps you stay clear on the bot's purpose as you build it, and it's what visitors will see when they chat.

Once you've named it, click Create and you'll be taken into the bot editor. Don't worry about all the options yet. We'll go through each one.

Step 2: Connect Your AI Brain

Right now, your chatbot is a shell — a structure with no intelligence inside it. This step is where it gets smart.

The intelligence comes from a large language model: Claude (made by Anthropic) or GPT-4 (made by OpenAI). These are the same AI models behind Claude.ai and ChatGPT. To connect them to your chatbot, you need an API key — think of it as a password that lets Botpress talk to the AI on your behalf.

To get a free Anthropic API key:

  1. Go to console.anthropic.com and create an account
  2. Click API Keys in the sidebar → Create Key → give it a name
  3. Copy the key immediately — Anthropic only shows it once. Paste it somewhere safe.

Then in Botpress:

  1. Go to Settings → Integrations → AI Models
  2. Paste your Anthropic key in the Anthropic section
  3. Select Claude as your active model and click Save

If you'd prefer GPT-4, the process is identical — just get your key from platform.openai.com instead. Not sure which to choose? The ChatGPT vs Claude 2026 comparison covers the differences. Either works well for a first chatbot.

Step 3: Give Your Chatbot a Personality and Purpose

This is the most important step that most tutorials rush past. The system prompt — the instructions you write that tell the AI who it is and what it should do — determines about 80% of how well your chatbot actually performs.

Think of the system prompt like a job description. The more specific you are about the role, responsibilities, and how the bot should behave in edge cases, the better it performs. A vague brief produces a vague chatbot.

Here's an example system prompt you can adapt for your own use:

You are a helpful assistant for [Business Name]. Your job is to answer questions about our products and services, help visitors find what they need, and collect their email if they want a callback. Stay friendly, concise, and honest. If you don't know the answer, say so and offer to connect them with a human.

When writing your own system prompt, think through four things:

Tone — How should the bot sound? Friendly and casual? Professional and precise? Match your brand voice. A law firm's chatbot should sound different from a creative studio's.

Scope — What topics should it answer, and what should it avoid? Be explicit: "Only answer questions about our accounting software. For everything else, politely redirect."

Uncertainty — What should it do when it doesn't know? The default AI behaviour is to guess, producing confident-sounding wrong answers. Fix this: "If you are unsure, say so and offer to connect the user with a human."

Data collection — Should the bot ask for contact details? Tell it when: "If a user asks about pricing or wants a demo, ask for their name and email first."

In Botpress, paste your system prompt into the Agent Instructions field in the bot editor. This is applied to every conversation.

Step 4: Add Your Knowledge Base

The system prompt tells your chatbot who it is. The knowledge base tells it what it knows.

A knowledge base is simply a collection of information your chatbot will draw from when answering questions. Upload your FAQs, your product descriptions, your pricing details, your return policy — anything you want the bot to be able to discuss accurately.

In Botpress, here's how to add it:

  1. Click Knowledge Base in the left sidebar
  2. Click Add Document
  3. You'll see three options: paste text directly, upload a PDF, or add a URL to scrape

For your first knowledge base, start by pasting text. Write out your 10 to 20 most commonly asked questions and their answers. Put the question on one line and the answer immediately below it. This format — clean Q&A pairs — gives the AI the clearest context to work from, and your bot will be noticeably more accurate than if you just dumped in a wall of text.

Once you've added your content, click Save and Sync. Botpress will process the document and make it available to the AI.

A common mistake: uploading a 50-page company handbook and expecting the bot to know everything in it perfectly. It won't. AI models work better with focused, specific information. Ten to twenty Q&A pairs will outperform a 50-page PDF every time. Start small and add more as you go.

Key tip: Treat your knowledge base like a living document, not a one-time upload. The questions real users ask will show you what's missing — and you'll add to it as you go.

Step 5: Test Your Chatbot

Before you show your chatbot to anyone else, test it thoroughly yourself. Botpress has a built-in Preview panel — click it and you'll see a chat interface where you can talk to your bot exactly as a real user would.

Work through these five test questions:

  1. A question that should be in your knowledge base — Does it answer accurately?
  2. A question that isn't — Does it admit uncertainty, or make something up?
  3. A follow-up question — Does it remember what was just discussed?
  4. An off-topic question — Does it stay focused, or wander?
  5. Something ambiguous — Does it ask for clarification, or guess?

Based on what you find, here's how to fix the most common problems:

Test until you're happy with all five categories. Don't skip this step and assume it'll be fine.

Step 6: Deploy and Embed

When you're satisfied with testing, it's time to put your chatbot somewhere real.

In Botpress, click Publish at the top of the editor. This makes your bot live. Then go to Integrations → Web Chat in the sidebar.

You'll see an embed snippet — a short piece of JavaScript code that Botpress generates automatically. It looks something like one line starting with <script. Copy it.

To add it to your website, paste that snippet just before the closing </body> tag in your HTML. That's the last line of code on your page. The chatbot widget will appear automatically in the bottom-right corner of every page that has the snippet.

If you use WordPress: go to Appearance → Theme Editor → footer.php and paste the snippet above the </body> tag. Alternatively, install the free plugin Insert Headers and Footers — it lets you paste the snippet without touching theme files.

If you don't have a website yet: Botpress also gives you a standalone shareable link. Go to Integrations → Shareable Link, copy the URL, and you can share your chatbot directly — as a link in an email, on social media, or in a message. It works instantly, no website required.

Making It Better: 5 Quick Improvements

Once your basic chatbot is live, these five changes will make it noticeably more useful:

  1. Add a welcome message that sets expectations from the start. Something like: "Hi, I'm Aria, Acme's assistant. Ask me anything about our products, pricing, or how to get started." This tells the user what the bot can do before they have to guess.

  2. Set a clear fallback response for when the bot can't answer. Rather than silence or an error, have it say something like: "I'm not sure about that one — want me to connect you with the team? Just leave your email and we'll get back to you." Turns a dead end into a lead.

  3. Add a lead capture step for high-intent questions. If someone asks about pricing or wants to book a demo, have the bot ask for their name and email before providing details. You'll be surprised how many people are happy to share it.

  4. Update your knowledge base regularly. Every time you get a question your bot couldn't handle well, add the answer to the knowledge base. This is the fastest way to improve accuracy over time.

  5. Check your conversation logs weekly. Botpress keeps a record of every conversation. Real user questions will show you gaps in your knowledge base and awkward phrasings in your system prompt faster than any amount of internal testing.

What's Next: Beyond the Basics

Once your chatbot is working and you've iterated on it a few times, you'll start to see what else is possible.

Multi-language support is built into Botpress — enable it in settings and the AI handles translation automatically. A five-minute upgrade if you have an international audience.

CRM integration — connect your chatbot to HubSpot, Notion, or a Google Sheet via Zapier. Leads flow directly into your existing workflow without manual work.

Voice — want your chatbot to speak rather than type? Connect ElevenLabs for realistic AI voiceovers. Our ElevenLabs review covers how it works and what it costs.

For automated pipelines where your chatbot triggers emails, updates databases, and runs without oversight, the guide to AI agents and automated workflows covers the next level.

Conclusion

You've built a working AI chatbot — given it an identity through the system prompt, knowledge through the knowledge base, and a home through the embed snippet. The same three elements apply to any AI tool you build on any platform.

Now let real people use it and pay attention to what they ask. That feedback is worth more than any further tinkering.

Share what you've built — and if you're just getting started with AI tools more broadly, the beginner's guide to getting started with AI is a good next read. For a broader look at which AI tools are worth your time across other parts of your work, the top AI tools for productivity in 2026 covers the full landscape.