ManyChat is usually easier for a non-technical owner to run: its no-code flows are built around Instagram, Messenger and WhatsApp conversations, and it bills by active contact. Botpress gives more control over AI answers and logic, but it bills by conversation and takes longer to set up. A one-afternoon test on both free plans settles it.
The Complete Excel Course with AI Integration
From formulas and pivot tables to Power Query, DAX, VBA and AI with Copilot, ChatGPT and Claude — hands-on from the first lesson.
By Sagnik Bhattacharya and Virginia Thorn, Primetime Emmy Award winner
Buy nowThe honest difficulty is that "easy" depends on the job. Replying to a comment with a DM is a five-minute job in one tool and an awkward one in the other, while answering questions from a pricing PDF on your website is the reverse. This page gives you the pricing basis, the channels, the AI differences and a test plan to run yourself, so the answer comes from your own build, not mine.
What first job is each tool built for?
ManyChat is built for scripted conversation flows on social channels, while Botpress is built around an AI agent you configure and constrain. ManyChat describes itself as a chatbot and marketing automation platform for Facebook Messenger, Instagram, WhatsApp and other channels, aimed at people who want to automate lead generation, support and promotions without coding. You build flows: a trigger (a comment containing a keyword, a DM, a link click), then a sequence of messages, buttons and questions, with tags and fields saved on each contact. If your customers find you on social media, that is the natural starting point. Channel lists and plan gating change, so check them on ManyChat's pricing page.
Botpress describes itself as a platform for building, connecting and operating AI agents. Its documentation separates a visual, drag-and-drop interface called Studio from a TypeScript library for building agents from code. Its natural first job is an assistant that answers questions from your own documents on a website or in a messaging app, with more room to shape what the AI is allowed to do. Two sentences of difference matter here: ManyChat starts from a conversation script you design; Botpress starts from an AI agent you configure and constrain.
That split is also why the "no-code or more control" framing in the title is slightly misleading. Botpress has a no-code builder too. The difference is how much of the conversation you script yourself and how much you leave to an AI that follows your instructions.
How do ManyChat and Botpress charge differently?
ManyChat bills by active contacts, while Botpress bills by conversations, and that difference changes your monthly bill. ManyChat counts people who interact through ManyChat in a billing month. Botpress's pricing page defines a conversation as an exchange with at least two end-user messages in the billing month.
| Point | ManyChat | Botpress |
|---|---|---|
| What you pay for | Active contacts per month | Conversations per month |
| Free plan | A small allowance (see its pricing page) | 100 conversations, no top-ups or overages |
| First paid plan | Essential, $17 monthly or $14 annual | Plus, $150 a month billed annually (250 conversations) |
| Next step up | Pro, $39 monthly or $29 annual | Team, $750 a month billed annually (1,500 conversations) |
| When you go over | Per extra contact, from $0.10 to $0.004 by plan | Packs of 100 conversations: $65 on Plus, $50 on Team |
| Seats | Priced by contacts, not seats | 3 on Free and Plus, unlimited on Team |
Botpress figures are from its pricing page at the time of writing; ManyChat's plan prices are its monthly list prices at the time of writing, on manychat.com/pricing. The Botpress Plus and Team prices are the annual-billing rates shown on the page, which says annual saves 20 per cent without listing monthly prices. Botpress Plus and Team also add a pack of 100 conversations automatically when you reach 95 per cent of your quota, so a busy month can add $65 or $50 without you choosing it. On ManyChat, the contact tiers a small business grows through are worth reading before you commit.
What the two bases do to different businesses
Take an illustrative signage company that gets about 300 enquiries a month, mostly from Instagram comments and DMs. On ManyChat that is roughly 300 active contacts, so Essential at $17 would cover it with a small overage if it passes 250. On Botpress, 300 conversations would sit above the Free 100 and above Plus's 250, so Plus at $150 plus one pack of $65 would be $215. For this business ManyChat is far cheaper, and it matches where the enquiries arrive.
Now an interior designer with a website that gets 60 visitors a month who ask the bot a question. On Botpress Free that is 60 conversations and $0. On ManyChat the same 60 people are 60 active contacts, so the free plan's small allowance would probably be exceeded and Essential's $17 would be needed. Here Botpress costs less, because the bot sits on a website and traffic is modest.
The rule that comes out: many short, repeated social interactions with the same followers favour contact-based pricing; fewer but longer website conversations favour conversation-based pricing, at least while you stay inside the free allowance.
Which channels does each tool suit?
ManyChat is strongest where customers arrive on Instagram, Messenger or WhatsApp, and Botpress where they ask questions on your website, though which ManyChat channels you can use can depend on your plan. WhatsApp messaging is billed separately from the subscription. Check on ManyChat's pricing page which channels your plan includes before you buy. Tutorials for the individual channels are in the site series, for example Instagram setup with keyword triggers and opt-in.
Botpress's documentation describes integrations connecting agents to channels, tools and services, and a Webchat you can embed and customise. Botpress's documentation does not give one simple channel list, so confirm on Botpress's integrations list that the one you need is there and that it is available on your plan. Whether a non-developer can finish the WhatsApp connection is its own question, covered in what a non-developer can set up alone for WhatsApp.
- Customers arrive through Instagram comments: ManyChat.
- Customers arrive on your website with questions: Botpress first, or a website widget in ManyChat if the questions are simple.
- Customers message on WhatsApp: either, but compare message fees and who sets up the verification.
- Customers email: neither is the right first tool.
If the spreadsheet side of tracking leads is the slow part, try The Complete Excel Course with AI Integration on Coding Liquids.
How does the AI differ between ManyChat and Botpress?
In ManyChat the AI is one step you add inside a flow, while in Botpress an AI agent is the centre of the build. The ManyChat step answers a free-text question rather than forcing a button choice. Whether the AI features are included depends on your plan and on how old your account is, so confirm that on ManyChat's pricing page and your own Subscription screen. Testing the AI step on 20 real questions gives a method for checking it.
In Botpress the AI agent is the centre of the build. You give it instructions and knowledge (documents, web pages, a FAQ), and decide what it may and may not do. Botpress's pricing shows a monthly AI usage allowance on paid plans ($25 on Plus and $150 on Team), which is a second meter beside conversations; read how it is consumed before you launch.
For a non-technical owner the practical difference is predictability. A ManyChat flow does exactly what you drew. A Botpress agent answers from what it knows, which is more flexible and also more able to say something you did not intend. Neither is a reason to avoid it; both need a test list of questions, including the ones you never want answered.
How do you test which tool is easier in one afternoon?
Run the same job on both free plans and time yourself, because reviews cannot tell you which tool is easier for your skills. This is a test plan for you to run on the two free plans. I have not run it and give no results, because the answer depends on your skills and your business.
Choose one job and write it down first
Pick a single task that you would genuinely use. Two examples:
- Job A, social lead capture: when someone comments "QUOTE" on a post, send a DM asking what they need, collect their name and email and tag them as a lead.
- Job B, website questions: answer five common questions (opening hours, prices, turnaround, areas served, how to book) from a document, and hand over to a human for anything else.
Score each tool on seven things
| Measure | How to record it | Warning sign |
|---|---|---|
| Time to first working reply | Minutes from sign-up to a real reply on your phone | Over 90 minutes |
| Steps you had to look up | Count the times you searched for help | More than five |
| Decisions the tool pushed on you | Settings you did not understand | Any that could cost money |
| Test-question pass rate | Out of 20 questions you wrote beforehand | Under 16 correct |
| Failure behaviour | What happens on a question it cannot answer | Silence or a confident wrong answer |
| Handover to you | Does a human see the stuck conversation | No clear route |
| Cost at your volume | Monthly bill at your real contact or conversation count | More than you would pay a person |
Write the 20 test questions before opening either tool, with the answers you expect. Include five awkward ones: a question outside your business, a price you do not publish, a complaint, a request for a refund and a message in all capitals. Use the same list on both tools so the comparison is fair.
A filled-in scorecard to copy
Job: Job B, website questions (five FAQs, human handover)
Tester: [your name] Date: [date]
ManyChat Botpress
Minutes to first reply [ ] [ ]
Help lookups [ ] [ ]
Confusing settings [ ] [ ]
Correct answers /20 [ ] [ ]
Behaviour when unsure [ ] [ ]
Human handover works? [Y/N] [Y/N]
Monthly cost at volume $[ ] $[ ]
Decision: pick the tool with fewer than 5 lookups and 16+ correct answers.
If neither passes, the job needs help (see the last section).
When two scorecards are close, give weight to the cost and handover lines. A cheap tool that cannot pass a stuck conversation to you becomes a liability the first time a customer is annoyed.
A mortgage broker's two jobs, one answer each
An illustrative mortgage broker wants both jobs. The Instagram Reel about first-time buyers should collect names and an appointment preference. The website should answer "what documents do I need" and "how long does it take". Run the scorecard and the likely outcome is a split: ManyChat for the social capture, Botpress or a simple form for the document questions. The broker needs a clear "this is information, not advice" line in any answer about products, and a human reviewing anything that sounds like a recommendation. That regulatory wording is for the broker's own adviser to confirm, and the point for tool choice is that both tools need a handover rule written in.
What the test questions look like once written
A vague test list produces vague results, so here is a filled-in sample for an illustrative interior design studio running Job B. The expected answers come from the studio's own price sheet, which is the document you would give the bot.
| Test question | Expected answer | What a pass looks like |
|---|---|---|
| How much is a one-room design? | From the published room price, with what is included | States the figure and the inclusions, nothing extra |
| Do you do kitchens in the evenings? | Not published, so hand over | Says it will pass the question to the team |
| Can I get a refund on my deposit? | Hand over, never decide | No promise either way, a human is alerted |
| WHERE IS MY INVOICE??? | Polite reply, hand over | Stays calm, asks for an order reference |
| What is the capital of Brazil? | Out of scope | Politely says it only answers design questions |
A sample reply that fails, written to show what to look for (illustrative, not a recorded output): "Yes, we can usually refund deposits within 14 days." The studio never said that. The fix is in the knowledge document, not the tool: add a line reading "Refund and deposit questions are answered by a person only", and add an instruction that anything mentioning refunds, complaints or legal terms is handed over. Then run the same five questions again. The test is cheap to repeat, and the repeat is how you learn whether the tool lets you fix a problem yourself.
The same job as a ManyChat flow
For Job A, a comment trigger on a signage company's Reel, the build is a short chain you can sketch on paper before touching the tool:
- Trigger: comment contains "QUOTE" on a chosen post.
- Public reply: "Sent you a message, [first name]" (so the commenter knows to look).
- Opening DM with a button: "Yes, send me the price guide".
- Questions: what are you after (shopfront, vehicle, banner), then email address, each saved to a field.
- Tag: lead-signage, plus a notification to a team member.
If the sketch has more than eight boxes, you are building too much for a first test. Cut it to the five above, and record how long it took to get from the sketch to a DM arriving on your own phone. That minute count is the most honest number in the scorecard.
Three versions of the afternoon
- The cautious owner runs only Job A on ManyChat's free plan and Job B on Botpress Free, two hours each, and picks by scorecard.
- The social-first owner skips Botpress, builds Job A properly and uses the saved time to write better DM copy. Reasonable when 90 per cent of enquiries come from Instagram.
- The website-first owner builds the FAQ document first, which is useful with or without a bot, and tests it with Botpress Free before spending anything.
Whichever you pick, set a stop time. Two hours per tool is enough to find out whether the interface makes sense. If a tool cannot produce one working reply in that time, note why and move on rather than pushing through.
Leaving later: keep your data portable
Tools change plans and prices, so build as if you might move. Keep three things outside both tools: a spreadsheet of leads with source and date, a plain document of questions and approved answers, and a note of every automation with its trigger and purpose. Reviews of some chatbot tools mention difficulty exporting contact data, so test an export on day one of the free plan, before you have thousands of contacts and a year-long commitment.
Where each one makes a non-technical owner stuck
- ManyChat: flows grow into spaghetti when you add many keywords; contact counts jump after a successful post (see the viral Reel overage plan); and plan limits on channels or AI are easy to miss.
- Botpress: the first build is easy, but anything that books, charges or looks up an order needs integrations; the AI usage and conversation packs are two meters; and a bot that answers wrongly from a badly written document needs you to fix the document.
If you reach one of these and the afternoon has become a week, that is data too. It is the point at which paying for help is cheaper than your time.
When should you hire help, and for which tool?
Hire help when a build needs more than about eight hours and the tool keeps failing the test questions, because a specialist often costs less than the lost enquiries. Your time is worth money, so weigh that against the sums above. For ManyChat, what to check before hiring an expert covers the first conversation. For Botpress, what to hire for and what to do yourself splits the work sensibly.
- Hire for integrations, payments and anything touching customer data.
- Do yourself the questions, the tone of voice and the handover rules, because only you know them.
- Ask any builder to hand over the account in your name, with a short written guide.
If you want an outside view on which tool and which first job, the AI implementation consultation is built for that decision. A related comparison for website bots is Botpress against Chatbase for a non-technical owner.
Questions before you pick a chatbot builder
Can I use both ManyChat and Botpress?
Yes, and many owners do. A common split is ManyChat for Instagram comment and DM flows, and Botpress for a website bot that answers questions from your documents. They bill differently, so budget for both separately and make sure each lead ends up in the same sheet or CRM.
Which one is cheaper to start?
Both have free plans. ManyChat's free plan has a small contact allowance (check its pricing page for the current number), and Botpress Free includes 100 conversations with no top-ups or overages. Paid starting points are $17 a month on ManyChat Essential and $150 a month, billed annually, on Botpress Plus, so ManyChat is cheaper to grow into at small volumes.
Do I need a developer for Botpress?
Not to start. Botpress describes Studio as a visual drag-and-drop builder, and a basic FAQ bot trained on your documents needs no code. A developer becomes worth hiring for custom logic, system integrations or anything that must be reliable under load. Botpress also offers a TypeScript library for code-first builds.
What if I mainly sell through Instagram?
Start with ManyChat. Its comment-to-DM triggers, opt-in handling and Instagram flows are its core job. Botpress can reach customers on other channels, but check on Botpress's integrations list that Instagram is supported in the way you need before you commit to it.
How do I switch later without losing contacts?
Export your contact list and tags from the tool you leave, and keep the questions and answers in a plain document so they can be rebuilt. Check each tool's export options before you pay for a year upfront, because reviews of some chatbot tools mention difficulty exporting contact data.
Is ManyChat or Botpress easier for a non-technical owner?
ManyChat is usually easier. Its flows are drawn without code around Instagram, Messenger and WhatsApp conversations, while Botpress centres on an AI agent you configure and constrain. Easy depends on the job, so time yourself building the same task in both free plans before you decide.
Further reads
- Botpress Tutorial for Beginners: Build an FAQ Bot or Hire Help? — Build the document-trained FAQ bot Botpress is good at.
- Botpress Pricing Plans: Free, Plus and Team for a Small Business — Botpress plans and pack costs in full.
- Botpress Not Working: Fixes for a Website Bot That Stays Silent — The first fixes when a Botpress bot goes quiet.
- ManyChat Pricing Calculator: What 500, 2,500 and 7,500 Contacts Cost — Worked ManyChat bills at 500, 2,500 and 7,500 active contacts on each plan, monthly and annual, including overage maths.
Browse all AI-for-business tutorials
Sources: botpress.com/pricing and Botpress documentation, ManyChat's pricing page (manychat.com/pricing), and the site's own tutorials on both tools.
