Full-stack platform for building and deploying LLM-powered AI agents
Botpress is an AI agent platform for developers and teams who want to build, deploy, and manage LLM-powered conversational agents.
AI Panel Score
6 AI reviews
Reviewed
In practice, users build agents inside Botpress Studio, a visual environment where workflows, prompts, and conversational logic are configured side by side. Agents can be connected to knowledge bases for RAG-based responses, assigned personality traits to match a brand voice, and extended with custom JavaScript actions for advanced automation. Once built, agents are deployed to a hosted cloud environment and distributed across channels including web chat widgets, WhatsApp, Slack, and Messenger.
Botpress distinguishes itself through several specific capabilities: an Autonomous mode that enables agents to use tools and make decisions across multi-step tasks; a Tables feature providing built-in structured storage for agent state and data; a Policy Agent that enforces content and safety rules to reduce hallucinations; and a Human-in-the-Loop (HITL) system that lets human agents take over live conversations with full context. A hub of prebuilt cards, workflows, and integrations allows teams to extend agent functionality without writing code from scratch.
Botpress targets developers, startup teams, and enterprise organizations building customer support automation, sales assistants, and internal workflow tools. The platform offers a free tier alongside paid subscription plans, with pricing scaling based on usage and features. It competes with platforms such as Voiceflow, Rasa, and Microsoft Copilot Studio in the conversational AI and agent-builder category.
Developers can access a REST API and JavaScript SDK to integrate Botpress agents into custom applications or backend systems. The platform is fully cloud-hosted, with data privacy and compliance details published on a dedicated trust page. Official integrations cover CRMs, databases, and major messaging platforms through both native connectors and API-based connections.
Enables agents to use tools and make decisions autonomously across multi-step tasks and outcomes without manual intervention.
Provides retrieval-augmented generation using embeddings and managed data sources so agents can answer questions from uploaded content.
Tracks agent performance, user engagement, and conversation quality including support for custom event tracking.
Allows developers to embed reusable JavaScript logic within agent workflows to handle advanced automation steps.
A visual builder that unifies prompt design, workflow creation, and conversational logic for building and configuring AI agents.
Built-in structured storage for managing agent state and operational data within the platform.
An embeddable, customizable chat interface that can be styled and deployed on websites or applications.
Controls the voice, tone, and behavioral style of each AI agent to match a specific brand identity or persona.
Connects agents to external systems such as WhatsApp, Slack, Messenger, CRMs, and databases through APIs and official connectors.
Provides a JavaScript SDK and REST API reference for integrating Botpress agents into custom applications or backend systems programmatically.
Enforces configurable constraints and safety rules on agent outputs to prevent hallucinations or unsafe responses.
Escalates live conversations to human agents with full conversation context, and allows humans to review or take over interactions in real time.
Free entry tier — 5,000 messages/month, 1 bot, 50 MB vector DB. Pay $0.01 per message beyond.
For growing builders — unlimited bots, 25% add-on savings, increased limits.
For teams — live agent handoff, real-time collaboration, RBAC.
Managed service — custom development, maintenance, monthly strategy calls.
Custom enterprise pricing with dedicated support. Contact sales.
Botpress did the LLM pivot without breaking its product — Series B closed, builders still ship on it.
“Botpress closed a $25M Series B in May 2025 at a $120M post-money, bringing total raised to $45M with 119 people in Quebec City. That's a defensible 36-month bet against Voiceflow's design-first ceiling and Microsoft Copilot Studio's enterprise lock-in.”
Most chatbot builders pivoted into LLM-agent platforms in 2023 and broke their products doing it. Botpress did the migration and still has a usable Studio on the other side. $25M Series B in May 2025, $120M post-money, 119 people in Quebec City.
Autonomous Node is what makes this credible — agents pick tools and make decisions without hand-coded flows, which is where Voiceflow still feels like 2022. Tables gives you structured agent memory without a side database. Policy Agent constrains outputs before you ship. Plus runs $89/month, Team $495 — both undercut Microsoft Copilot Studio on real per-message economics.
The catch is the meter. Pay-as-you-go is $0.01 per message past 5,000, and Always Alive plus knowledge-base storage stack on top once volume picks up. Pilot one bot on Plus for a quarter before standardizing across support.
Credible alternative to Voiceflow and Microsoft Copilot Studio with sharper LLM-native primitives.
HubSpot Ventures and Deloitte Ventures on the cap table make the choice defensible to a board.
Visual Studio plus RAG plus prebuilt cards enable fast pilots, but multi-channel deployment still takes work.
LLM-agent platform with Autonomous Node moves teams forward, not just cost-saving on existing chatbots.
Series B closed May 2025 at $120M post-money, $45M total raised, 119 people, 9 years in market.
Teams who need an LLM-agent platform with structured memory and policy controls.
Solo builders who only need a static FAQ chatbot.
Botpress turns the Policy Agent into the production seam most agent platforms forgot to ship.
“Founded 2017 in Quebec by Sylvain Perron, Botpress closed a $25M Series B in June 2025 led by Framework with HubSpot Ventures participating. For a CTO picking the conversational-agent stack through 2029, the call is whether a JavaScript-extensible runtime with a Policy Agent beats Voiceflow's visual ceiling or Rasa's open-source floor.”
Most agent platforms ship orchestration and skip the guardrails. Botpress puts a Policy Agent in the runtime to enforce output constraints — the piece a CTO has to defend in an incident review. Knowledge Bases handle RAG and the Autonomous Engine handles tool use, so the surface reads production-shaped, not demo-shaped.
Plus at $89/month opens unlimited bots; Team at $495 unlocks live agent handoff and RBAC, the seam regulated workloads land on. Founded 2017 in Quebec by Sylvain Perron, Botpress closed a $25M Series B in June 2025 led by Framework with HubSpot Ventures participating. Against Voiceflow's design-led shape and Rasa's $35K/year self-hosted floor, it sits as the JavaScript-extensible middle.
But the catch is runtime lock-in — Custom Actions and Tables live inside their stack, not yours. For a CTO picking the agent platform through 2029, the per-message meter scales cleanly until volume hits the point where Rasa's annual contract pencils out. Strong default bet underneath that ceiling.
Sits cleanly between Voiceflow's visual ceiling and Rasa's self-hosted floor — the JavaScript middle.
JavaScript SDK with Custom Actions matches how engineering teams extend conversational runtimes.
WhatsApp, Slack, Messenger, HubSpot, and Notion connectors plus REST API cover the standard stack.
Runtime lock-in through Tables and Custom Actions is real, but the per-message meter delays the rebuild question.
Policy Agent plus Autonomous Engine plus RAG-grade Knowledge Bases is production craft, not demo craft.
CTOs who need a production runtime for LLM agents.
Teams who want pure code-only orchestration without a managed runtime.
Three published tiers, but every one rides on '+ AI spend' — the invoice you can't model.
“Plus is $89/month, Team $495, Managed $1,245 — all surcharged with AI spend passed through at OpenAI and Anthropic rates. Free Pay-as-You-Go covers 5,000 messages with $0.01/message overage, a defensible floor for pilots but not for procurement modeling.”
The line item that matters isn't on the pricing page. "+ AI spend" sits next to every paid tier. Plus is $89/month. Team $495. Managed $1,245. The model rate is whatever OpenAI or Anthropic bills that quarter. Procurement can't lock it.
Pay-as-You-Go covers 5,000 messages/month, 1 bot, then $0.01/message. A team running 50K messages pays $450/month before AI spend. Plus at $89/month adds unlimited bots and 25% off add-ons. Team at $495/month adds RBAC and live handoff. Compare Voiceflow Pro at $50/editor and Intercom's Fin at $0.99 per resolution — different unit, similar drift.
Founded 2017 in Quebec. Raised a $25M Series B in June 2025 led by Framework Venture Partners. Always Alive is now bundled, no cold-start tax. The catch is invoice predictability — model pass-through means the bill moves with token prices. Honest platform fee, unpredictable AI line.
Variable AI line item is procurement friction; SSO and volume discounts gated behind Enterprise sales call.
Monthly billing on Plus, Team, and Managed; no public auto-renewal trap and unlimited bots included on every paid tier.
Three paid tiers fully visible on the pricing page, but "+ AI spend" line item obscures the real monthly bill.
Per-message economics at $0.01 give a clean unit-cost view; 5,000 free messages let teams measure resolution rate before committing.
Model pass-through pricing means token-rate moves directly hit the invoice; 3-year cost depends on OpenAI and Anthropic list prices.
Teams who pilot AI agents on the free tier before scaling.
Buyers who need a fixed monthly bill without model pass-through.
Policy Agent and Autonomous mode let a chatbot engineer ship multi-step agents without rewriting the guardrails.
“Policy Agent enforces safety rules at the output layer while Autonomous mode handles multi-step tool use, so an engineer doesn't stack a separate guardrail library on top. But the 5,000-message free tier vanishes the first week a production widget catches real traffic, and overage at $0.01 per message adds up before the Plus tier covers it.”
Building a production chatbot means two parallel fights — making the agent actually do the multi-step thing, and stopping it from saying something wrong. Botpress Studio puts both in one workspace. Autonomous mode handles tool-using decisions across multi-step tasks; Policy Agent enforces output constraints in the same place.
Tables ships built-in structured storage, so agent state doesn't need a side-car Redis. HITL escalation carries full conversation context to the human queue. The JavaScript SDK and REST API mean custom actions slot into the workflow without leaving the editor. Voiceflow nails the design surface but punts on the storage layer.
But the credit math bites — 5,000 messages monthly on free, then $0.01 each, and Live Agent Handoff only unlocks at Team at $495 a month. $25M Series B in June 2025 from FRAMEWORK signals runway, not feature-completeness.
Studio, Policy Agent, and Tables reduce the side-car infrastructure an engineer normally bolts around an LLM agent.
Pricing page, docs, changelog, and a developer blog all ship live with concrete tier details.
Credit math on the free tier plus HITL gated to the $495/mo Team plan adds real weekly friction.
Custom JavaScript Actions plus Autonomous mode plus the integration hub scale cleanly from prototype to production.
JavaScript SDK plus REST API plus native WhatsApp/Slack/Messenger channels match how dev teams already ship.
Chatbot engineers who ship LLM-powered agents to multiple channels.
Hobbyists who need more than 5,000 messages on free.
Botpress is what an agent builder looks like when the team has actually shipped chatbots before.
“Botpress pairs a visual studio with autonomous tool-use, RAG knowledge bases, and a Policy Agent that keeps outputs in their lane. Free tier covers 5,000 messages/month and Plus is $89/month — solid pricing for a 2017-era platform with a fresh $25M Series B.”
Most bot builders show you a canvas and pray you figure out what goes where. Botpress Studio at least labels the lanes — Knowledge Bases on one side, Tables for state, the Policy Agent gatekeeping the output. Whoever drew the layout has built a bot before.
Pay-as-you-go is 5,000 messages free, $0.01 each after. Plus is $89/month, Team jumps to $495 for live handoff and RBAC. Always Alive — their no-cold-start feature — landed on every paid plan in May 2026, which reads like a vendor who actually listened. Founded 2017 in Quebec, $25M Series B closed last summer.
But Autonomous mode is the kind of feature that demos beautifully and breaks expensively if you skip the guardrails. Voiceflow is friendlier to non-developers; Microsoft Copilot Studio buys you the M365 hug. Botpress is the one you pick when you want JavaScript actions instead of a coloring book.
Botpress Studio labels lanes clearly — Knowledge Bases, Tables, Policy Agent — a layout drawn by people who have built a bot before.
Visual builder is approachable at first; Autonomous mode and JavaScript Custom Actions raise the ceiling and the ramp.
Dev infrastructure for building agents — mobile parity is not the relevant use case, scored neutral.
Pay-as-You-Go with 5,000 free messages lowers the entry bar, but the agent paradigm itself takes time to click.
Always Alive (cold-start removal) and HITL with full context suggest a team that thinks about live-traffic edge cases.
Developer-leaning teams who need autonomous agents with real guardrails.
Non-technical operators who want a no-code chatbot in an afternoon.
Botpress is the open-source chatbot project that actually re-platformed for the LLM agent era.
“Founded 2017 in Quebec by Sylvain Perron, Botpress closed a $25M Series B from Framework Venture Partners in June 2025 at a $120M post-money valuation, then rebuilt around Autonomous mode and the Policy Agent. The catch is the field — Voiceflow, Microsoft Copilot Studio, and a fresh wave of agent-frameworks all want the same buyer.”
Watch how a chatbot vendor handles the LLM transition. Botpress re-platformed instead of pivoting away. Open-source roots from 2017, Quebec team under Sylvain Perron, $25M Series B led by Framework Venture Partners in June 2025 at a $120M post-money valuation. Cap table earned the runway.
The Studio is real work. Autonomous mode, the Policy Agent, Tables for state. Pay-as-You-Go is genuinely free at 5,000 messages, Plus at $89, Team at $495 with HITL and RBAC. JavaScript SDK and REST API make it look more like infrastructure than a no-code toy.
But the category is loud. Voiceflow has the design-tool framing, Microsoft Copilot Studio owns the M365 footprint, every LangChain wrapper claims agent-builder. Open-source heritage is the moat Botpress hasn't fully pressed. Exit is decent — JS actions are portable, RAG content yours. Worth the Plus tier, hedged on category compression.
Open-source heritage and Policy Agent help, but Voiceflow and Microsoft Copilot Studio crowd the same buyer.
JavaScript actions and RAG content are portable, but Studio workflow logic is harder to lift off cleanly.
$25M Series B from Framework Venture Partners in June 2025 at $120M post-money buys real runway.
Complete AI Agent Platform framing matches what Studio, Autonomous mode, and the Policy Agent actually deliver.
Open-source from 2017 then re-platformed for LLM agents — a survivor pattern most pre-LLM chatbot vendors missed.
Developer-led teams who build LLM agents.
Solo no-code builders who need ready-made templates.
Common questions answered by our AI research team
Pay-as-you-go is free (5,000 messages/month, 1 bot). Plus $89/mo, Team $495/mo, Managed $1,245/mo, Enterprise custom. All paid plans include unlimited bots.
Yes. Pay-as-You-Go includes 5,000 messages/events per month, 1 bot, 50 MB vector database, and 1 collaborator. Beyond that it's $0.01 per message.
Yes. Live agent handoff and real-time collaboration are introduced in the Team plan ($495/mo), along with RBAC and enhanced knowledge bases.
Always Alive keeps bots responding instantly from the first message — no cold-start latency. Since May 2026 it's included on every paid plan at no extra cost.
Managed at $1,245/month adds custom bot development by Botpress, ongoing maintenance to reduce model drift, and monthly strategy calls with their team.
Botpress is a Montreal-based platform for building and deploying AI-powered chatbots and conversational agents, built on top of large language models.
