9 of the best no-code chatbot builders

This guide is published by Figma. Figma Make is included in this roundup as a Figma product.

Your team wants a chatbot on the website by the end of the quarter, and nobody on the team writes code. Designers and PMs run into this problem constantly. A no-code chatbot builder means you can start without a developer. Its visual canvas lets you map what the bot says and how it recovers when someone goes off script.

In our 2026 AI report, 83% of respondents said they’re working on AI-powered projects, up from 65% a year earlier. Chatbots are a common starting point, and the tools for building one vary from simple website widgets to full chatbot platforms with LLM routing and CRM connections.

Choosing the right builder depends on what your bot has to do and how much of the conversation you want to shape.

Read on to learn:

• Nine of the best chatbot builders and what each one does well
• How to choose the right chatbot platform for your team
• When no-code is enough and when you need a developer
• Answers to common questions about chatbot software

1. Figma Make

Ideal for: Prototyping chatbot conversations and states before you build

Figma Make turns a plain-language prompt into a working chatbot prototype. Describe the logic and personality you want, then run a real conversation through it.

Because your chatbot runs on working code, you can pressure-test how it behaves before handoff. That includes the moments when things go wrong, like a slow reply or a failed request.

Form is worth testing, too. A full-page assistant, a docked panel, and an inline copilot are all different products, even with identical logic underneath. Prompt each one, compare them side by side, then refine the winner on the canvas or in the code, whichever is quicker.

Teams with a React design system can bring it in through Make kits or npm packages, giving your prototypes a set of components and tokens to work from. Figma Make also connects to external tools through Model Context Protocol (MCP), which lets documents and data from other apps feed the bot as real context.

Key features

• Prompt-to-prototype generation. Write out how the bot should handle a request and get an interactive app you can open and use.
• Design system support. Import your React library through a Make kit or an npm package, so prompts start from your system.
• State and behavior testing. Simulate branching paths and empty states to see how the bot holds up when a conversation goes off script.
• MCP connections. Pull documents and records from tools your team already runs into the prototype as context.
• Code and canvas together. Select any element and adjust it in the properties panel, or leave an in-context annotation describing what should change.

2. Botpress

Ideal for: Building AI agents that follow strict logic and call your APIs

Botpress is an AI chatbot builder where you lay out conversation logic as connected nodes on a canvas. It’s a good fit when a user flow has to behave the same way every time, like an eligibility check or a guided booking sequence. Drop in an Autonomous Node, and a model takes over the open-ended stretches while your scripted paths stay locked down.

Knowledge Bases supply the source material. Upload PDFs or Word docs, or connect a Botpress Table so the agent answers from structured records. From there, an agent deploys to WhatsApp, Instagram, Facebook Messenger, and Slack. Botpress Desk handles escalation when someone needs a person.

Key features

• Model flexibility. Run an agent on OpenAI, Anthropic, Groq, and more.
• Tables. Conversation data lives inside the agent, and lookups pull a record mid-chat without an outside database.
• Prebuilt agents. Add a translator or vision agent to handle a specific job inside a larger flow.
• Integration hub. Connect to HubSpot, Notion, Jira, and Calendly, plus custom code for anything not covered.

3. Voiceflow

Ideal for: Designing and testing conversation flows with your team

Voiceflow is a chatbot development tool built around a shared canvas, where a designer and a support lead can work on the same agent at once.

You build the fixed paths as workflows, dragging steps into the correct order so a return or a refund runs the same way every time. Anything harder to script becomes a playbook, where you write instructions in plain language and let the agent reason its way through.

Test an agent’s answers in a prototype before a customer ever sees it, then adjust what the knowledge base returns. After launch, transcripts and analytics show how the conversations actually went.

Voiceflow runs on chat and voice, and its Tools layer reaches APIs and MCP servers when an agent needs to do something outside the conversation.

Key features

• Knowledge base grounding. Voiceflow chunks and embeds your sources, with no vector database to stand up on your side.
• Platform APIs. Pull dialog state, knowledge management, and transcripts into your own custom interfaces.
• Tool calling. An agent can hit a CRM, a help desk, or an internal service mid-conversation and bring the answer back.
• Evals. Run tests against an agent to catch a regression before it reaches customers.

4. Landbot

Ideal for: Turning static forms into conversations that capture leads

In Landbot, you build a chatbot by dragging blocks onto a canvas and connecting them with arrows. Each block does one job, like sending a message or checking a condition, and the path you draw becomes the conversation.

Where it lands is up to you. A pop-up or widget works on an existing page, or the bot can take over a full landing page and replace your contact form outright.

Scripted blocks work well for a quoting or booking sequence that has to run the same way every time. When someone asks a question you never mapped, an AI Agent picks it up and answers in natural language before handing the conversation back to the flow.

A shared inbox lets a person take over when neither one is enough. Landbot also deploys to WhatsApp and Facebook Messenger, with WhatsApp running through the official Business API.

Key features

• Campaign manager. Send segmented outbound WhatsApp messages, then route the replies back into a flow.
• Native integrations. HubSpot, Airtable, Zapier, and webhooks push captured data into the tools your team already runs.
• A/B testing. Split traffic between two versions of a flow to see which one holds attention longer.
• Rich message blocks. Carousels and date pickers keep a conversation from turning into a wall of text.

5. Tidio

Ideal for: Adding live chat and support automation to an online store

With Tidio, you set rules for when a chat should open, like when a visitor lingers on a checkout page. Lyro, the AI agent, answers open questions, and a live agent handles whatever Lyro can’t resolve.

Lyro draws on your own material to answer a question, like your website or files you upload. It also reads three hours of recent history, so a visitor who was mid-conversation with an agent doesn't repeat themselves. The same setup runs across your site widget, WhatsApp, Instagram, and Facebook Messenger.

Key features

• Product recommendations. Lyro can suggest items from a Shopify or WooCommerce catalog, or from a product feed you upload.
• Copilot. Live agents get AI-written reply suggestions and conversation summaries while they’re in the chat.
• Lyro analytics. Answer rates and resolved counts show what’s working, and a list flags questions Lyro couldn’t handle.
• Multilingual support. Lyro responds in the visitor’s language without a separate bot for each market.

6. Manychat

Ideal for: Automating DMs and comment replies on social channels

Manychat runs on triggers. A comment containing a certain keyword or a Story mention kicks off an automation. The Flow Builder canvas helps you decide what happens next. Branches split based on how someone replies, so a person asking about sizing and a person asking about shipping end up on different paths.

Not every question fits a scripted branch. An AI Step takes a goal and some context about your business, then handles that stretch of conversation on its own. Everything runs across Instagram, Facebook Messenger, WhatsApp, TikTok, and Telegram, with a shared inbox for the conversations that need a person.

Key features

• Quick Automations. Guided setup walks you through a common automation without opening the full canvas.
• Comment-to-DM. A public comment on a post triggers a private message, so a lead doesn’t stall in the feed.
• Instagram Custom Audiences. Build first-party lists from people who interacted with your account, then retarget them in Meta Ads Manager.
• Smart delays. Space out messages so a sequence lands like a natural conversation.

7. Chatbase

Ideal for: Training a support agent on your own documentation

Chatbase builds an agent out of material you already have. Upload files, crawl your site, or connect a Notion page, then train the agent on those sources. The Playground lets you chat with the result and adjust its instructions before anyone else sees it.

Actions give the agent something to do besides answer. Built-in ones book a meeting through Calendly or collect a lead. A Custom Action connects to any API you run, so a customer request to downgrade a plan can happen inside the chat. One configuration deploys to a website widget, WhatsApp, Instagram, Facebook Messenger, Slack, and email.

Key features

• Model choice. Run an agent on OpenAI, Anthropic, Google, or DeepSeek, and set a different model per agent.
• Shopify actions. The agent syncs your product catalog and keeps up with changes, so it can surface items or check an order mid-chat.
• Topics and sentiment. Analytics group conversations by subject and flag how people felt, with exports for the rest of the team.
• White label. Host the agent on your own domain with Chatbase branding removed.

8. Intercom Fin

Ideal for: Resolving support questions inside an existing help desk

Fin is Intercom’s AI agent, and it works inside the same help desk your team already uses to answer tickets. It trains on your help center articles plus any documents or public URLs you add, then answers across the Intercom Messenger, email, WhatsApp, SMS, and Instagram.

Before Fin goes live, run simulations against past conversations and preview how a threaded email exchange will read. The Fin AI Engine refines each query and validates the answer before it reaches a customer. Anything Fin can’t resolve passes to one of your teammates. From there, Copilot drafts a reply and translates the conversation in real time.

Key features

• Fin Voice. Fin takes phone calls and connects the caller to a person when a question needs one.
• Image understanding. A customer can send a screenshot over chat or email, and Fin reads what’s in it before replying.
• Roles. Fin shifts between service and sales depending on what the conversation calls for.
• Multilingual replies. A conversation stays in the customer’s language from the first question through to the closing message.

9. HubSpot Chatbot Builder

Ideal for: Qualifying leads and booking meetings from your website

In HubSpot, a chatbot lives inside a chatflow on your site. The rule-based builder uses if/then branches to triage an incoming question, so a pricing request routes to sales, while a login problem becomes a support ticket. The widget reads from your CRM as it goes, shaping the conversation around what you already know about a contact. It then writes every exchange back to their record.

For questions a script can’t cover, the Breeze customer agent answers from your HubSpot content and any files or public URLs you add. When its confidence is low, it asks a follow-up question or reassigns the conversation to a rep. You set the rules for when that handoff happens.

Key features

• Meeting booking. A qualified visitor picks a time on a rep’s calendar without leaving the chat.
• Actions. The customer agent calls an external API mid-conversation to check an order or pull account data.
• Ticket creation. The chatbot opens a support ticket and automatically routes it to the right team.
• Language coverage. The customer agent works in every language HubSpot supports.

How to choose the right no-code chatbot builder

These tools don’t all cover the same ground. Some run a chatbot on your website, while others work inside social DMs or a help desk. The choice matters more six months in, when you need a new data source, or you want the same bot answering on WhatsApp.

Here are five things to check before you commit:

• Training sources. Check which models a tool supports and what it accepts as source material. If your content lives in a Notion workspace or a live database, check whether that connection exists before you commit.
• Channel coverage. List the channels you need, including email or WhatsApp if your team handles support there. Then ask whether one setup reaches all of them.
• Design flexibility. Some tools give you a color picker and a logo slot. Others let you shape the whole surface, including how it handles a slow response.
• Integrations. A bot that checks an order or books a meeting depends on native connections to your stack or an API you can call.
• Data handling. Check each vendor’s terms on where conversations are stored, how long they’re kept, and whether they’re used for model training. Settle this early if you work in a regulated industry.

At some point, you might wonder whether no-code is enough. The answer depends on what you’re building. Custom retrieval logic or an interface built from scratch pushes you toward low-code territory or a developer with AI coding tools. A good middle path is to prototype the experience first, then hand a working reference to engineering.

The best chatbot builder for your team is the one that fits how you already work. A marketing team living in Instagram DMs has different requirements than a support team inside a help desk, and a bot that sells products looks different from one that answers policy questions.

Chatbot development FAQ

What is a no-code chatbot builder?

A no-code chatbot builder lets you create a working chatbot through a visual interface, with no programming required. You drag steps onto a canvas, connect them, and the path you build becomes the conversation. Most chatbot development tools in this category also let an AI model field the open-ended questions your scripted steps don’t cover.

What makes a good chatbot for a website?

The best chatbot for a website matches what the website does. A store needs product lookups and order status, while a documentation site needs accurate retrieval and a clean handoff when the bot doesn’t know an answer. How the widget looks and behaves matters too, since it’s part of the experience people have with your brand.

Can you build a chatbot with AI in Figma?

Yes. Figma Make turns a plain-language prompt into a working chatbot prototype you can hold a conversation with, built from your team’s own components. Test how it handles a slow response or a failed request, compare a docked panel against a full-page assistant, then hand the working version to engineering.

Build your chatbot in Figma Make

Picking a no-code chatbot builder gets easier once you know what your bot has to do and where it needs to live. The right one depends on how your team already works. Figma Make lets you shape the conversation design first, with an interactive version of your chatbot that draws on your own components and tokens. Run real conversations through it and see how it holds up before handoff.

Here’s how Figma can help you put it all together:

• Map the paths a user takes before you build them, using a FigJam whiteboard to work through the branches with your team.
• Pull your chat components into Figma Design so the widget matches the rest of your product.
• Hand the finished interface to engineering through Dev Mode, where developers can inspect specs and pull assets in context.
• Present the working prototype to stakeholders with Figma Slides and collect feedback in the same file.