Botpress Review (2026)
A visual builder for people who want to see and control exactly what the bot will do, rather than trusting a model to work it out.
Some links on this page can be affiliate links. We may earn a commission at no cost to you. This never affects our rankings. How we work.
Visit Botpress ↗Opens botpress.com in a new tab.
What the score measures. Five things we can verify from published material: whether pricing is transparent, whether the product can act on your systems or only answer from documents, whether the vendor still owns its own roadmap, how much independent review evidence exists, and whether the bill stays predictable as volume grows. Weighted, then scored against an ideal platform that scores 5.
BEST FOR: BUILDERS WHO WANT DETERMINISTIC FLOWS RATHER THAN A BLACK BOX · LAST UPDATED 2026-08-11
Botpress in depth
Botpress predates the current wave and has adapted rather than pivoted. Where newer tools ask you to point an agent at your documentation and trust it, Botpress asks you to draw the conversation.
That appeals to a specific builder: someone who wants to know what happens at every branch, who needs certain paths to be guaranteed rather than probable, and who is comfortable with a flow editor. The LLM handles understanding what the user meant; your flow decides what happens next.
What it does
- Visual flow builder with explicit branching
- Open-source core with a hosted cloud option
- LLM-backed understanding inside deterministic flows
- Multi-channel deployment: web, WhatsApp, Slack, Messenger
- Custom code steps where flows are not enough
- Usage-based cloud pricing with a free tier
More detail on how it works
It has an open-source core and a hosted cloud with usage-based pricing and a free tier, which makes it one of the few tools here you can evaluate seriously without talking to anyone.
The flow builder
Explicit, visual and inspectable. When a conversation goes wrong you can see which node did it, which is a real advantage over debugging a prompt. The cost is that every path you want has to be built.
LLM inside deterministic structure
The pragmatic middle position: model for comprehension, flow for control. Slower to build than pointing an agent at a docs site, and far more predictable in what it will do.
Open source and self-hosting
The core is open source, so you can self-host for data residency or run it without a vendor relationship. Cloud usage pricing has a free tier that is genuinely usable for evaluation.
Setting Botpress up
Realistic time to a working deployment: Hours for answers, weeks for actions. These are the standard steps for this category. We have not published a walkthrough specific to this tool yet.
- 01
Connect your content
Point the agent at your help centre, docs site or past tickets. Usually a URL and an OAuth grant. Minutes.
- 02
Let it index
The platform crawls and embeds your content. Anywhere from minutes to a few hours depending on volume.
- 03
Set the confidence threshold
Decide how sure the agent must be before it answers rather than escalating. This single setting drives both your resolution rate and your bill.
- 04
Test against real past tickets
Replay questions you have already answered and compare. The most valuable hour of the whole evaluation, and the one most often skipped.
- 05
Build custom actions
If you need order lookups, refunds or account changes, this is where a developer connects your APIs. Days to weeks.
- 06
Define the escalation path
Where a failed conversation lands, and whether the transcript travels with it. Broken handoff is the single most common complaint about deployed AI support.
- 07
Roll out to a slice of traffic
One channel or one ticket type first. Watch escalation reasons rather than deflection rate.
The step that takes longer than they imply. Custom actions. Anything that queries your own systems is engineering work, not configuration, and it is the step teams under-budget most.
What the AI actually does
Every platform here says “AI agent”. It covers a bot that reads your help centre and a system that can refund a customer. This is which one you are buying.
Botpress sells on exactly the distinction this section exists to draw.
Its own headline is that most support AI answers and Botpress resolves, and the capabilities it names are refunds, account updates and order corrections rather than answers.
- Named actions are issuing refunds, changing a plan, correcting billing details, reissuing a shipment and correcting an address
- Coordinates several systems in one motion — updates the record, triggers the next step, closes the loop — rather than one lookup per conversation
- Autonomous nodes let the agent decide when to follow a fixed workflow and when to reason, instead of committing to one or the other at build time
- Handoff carries the conversation history, account details and troubleshooting already attempted, and the agent stays available behind the human
- No per-seat cost and AI usage included in the price, which is the opposite of the second-meter model Gorgias and Fin bill on
Who Should Use It
The second list is the more useful one.
Builders who want deterministic behaviour
If certain paths must be guaranteed — a compliance disclosure, a specific escalation — a flow gives you that and a prompt does not.
Teams with an unusual process
When your support flow does not look like anyone else's, a builder beats a product with opinions about how support works.
Anyone needing self-hosting
The open-source core makes data residency achievable without a vendor negotiation.
Who should look elsewhere. Teams wanting to deploy this week. Botpress is a builder, and building takes time — if your questions are documented and ordinary, Chatbase or My AskAI will be live in an hour and answer them just as well. Also not the tool for autonomous action across business systems; that is a different product category.
Strengths and Weaknesses
What Works
- Open-source core with self-hosting available.
- Free tier you can actually evaluate with.
What Does Not
- Building flows is work, and it never stops being work as your cases grow.
- Usage-based cloud billing is exposed to traffic spikes.
- Not an autonomous agent — it does what you drew and nothing more.
Botpress pricing
Read from botpress.com/pricing (vendor's own), read in-browser on 2026-08-11.
$0
Pay-as-you-go AI usage
$150
/mo$25 of AI usage included, then $65 per 100 conversations, about $0.65 each
$750
/mo$150 of AI usage included, then $50 per 100 conversations, about $0.50 each. The rate falls as you commit, which is the opposite shape to most of this table
$40
/moEach adds capacity such as table rows and vector storage
Verified 2026-08-11 · botpress.com/pricing (vendor's own), read in-browser
Botpress Cloud is usage-based with a free tier, and the open-source core is free to self-host.
Usage billing means a traffic spike is a billing event, so check the rate against your peak rather than your average. Self-hosting removes that risk and adds the ongoing cost of running it.
The full arithmetic
Compared against Chatbase at around $40 a month, Botpress asks for more building effort in exchange for far more control. Which is the better deal depends entirely on whether you want to build or to deploy.
INTEGRATIONS
Our Recommendation
Who we would tell to buy this, and who we would not.
Botpress is the right choice when you want to control the conversation rather than delegate it. The visual builder makes behaviour inspectable and debuggable in a way prompt-driven agents are not, and the open-source core plus free tier means you can evaluate it properly without a sales call. The trade is effort: every path you want, you build. If your support questions are documented and conventional, a retrieval agent will get you further faster. If they are not, or if certain paths must be guaranteed, this is the more honest tool.
Similar Tools
Others in ai agents & chatbots.
Fin (formerly Intercom)
Teams already on Fin who can absorb variable per-resolution billing
4.2 / 5 · $0.99/resDecagon
Large support organisations that need the agent shaped to existing procedures
3.1 / 5 · Not publishedSierra
Enterprises wanting one vendor across chat and phone
3.1 / 5 · Not published
Ada
Global brands needing many languages and channels at once
3.5 / 5 · Not publishedCompared Head to Head
Same category, so the comparison is a real decision rather than a keyword.
Frequently Asked
The questions people actually search for about Botpress.
Is Botpress free?
The open-source core is free to self-host, and Botpress Cloud has a free tier with usage-based paid plans above it.
What is Botpress used for?
Building chatbots and support agents with a visual flow builder, where an LLM handles understanding and your flow controls what actually happens.
Is Botpress better than Chatbase?
More control and more work. Chatbase is live in an hour on your existing documentation; Botpress gives you deterministic behaviour if you build it. Different tools for different problems.
Can Botpress be self-hosted?
Yes, the core is open source, which makes it a real option where data residency rules out a hosted vendor.
Does Botpress use AI?
Yes, for understanding what the user means. The difference is that your flow decides what happens next rather than the model.
Can Botpress take actions in other systems?
Through custom code steps and API calls, yes, but it is a builder rather than an autonomous agent. Expect to write the integration.
What are the main Botpress alternatives?
Chatbase or Wonderchat for fast retrieval-based deployment, Rasa for another open-source route, Fin or Ada if you want autonomy rather than control.