# Botpress — An open-source (MIT), developer-first platform for building conversational AI agents with bi-directional MCP support — self-host free, or use the managed cloud from $0/month.

> Source: The Agents Index — https://theagentsindex.com/botpress (structured, researched, re-verified)
> Facts last verified: 2026-08-02

Botpress is an open-source (MIT-licensed) platform for building and deploying LLM-powered conversational AI agents, founded in 2017 in Quebec City, Canada by Sylvain Perron and Justin Watson. What began as a developer-first open-source chatbot framework has grown into a combined open-source core plus managed cloud product, backed by a $25M USD Series B (June 2025, $120M post-money valuation) from Framework Venture Partners, Deloitte Ventures, HubSpot Ventures, Inovia Capital and Decibel Partners. Its GitHub repository carries roughly 14.8k stars, and the company reports around 3,000 paying customers alongside nearly 1 million free-tier developers.

| Fact | Value |
| --- | --- |
| Website | https://botpress.com |
| Pricing | Free and open-source (MIT) to self-host indefinitely. The managed cloud starts at $0/month (pay-as-you-go, ~500 messages/month) then Plus $89/month and Team $495/month, both plus separate LLM "AI Spend" billing — Enterprise is custom. Figures are sourced from independent reviews, not Botpress's own pricing page, which returned no accessible content at verification time. |
| API | Yes |
| Best for | Engineering-led teams building a multi-channel conversational agent who want an open-source, self-hostable core with an optional managed cloud layer. |
| Not for | Non-technical teams that want a fully no-code, zero-engineering chatbot builder — Botsify or a similarly managed platform is a better fit. |

## Pricing

| Tier | Price |
| --- | --- |
| Pay-as-you-go | Free |
| Plus | $89 / month + AI Spend |
| Team | $495 / month + AI Spend |
| Enterprise | Custom |
| Self-hosted (open source) | Free |

## Verdict

Botpress is the pick for engineering-led teams that want to own a conversational agent's code and data rather than configure a closed, fully-managed builder: the core framework is genuinely open-source (MIT-licensed, ~14.8k GitHub stars) and free to self-host indefinitely, with bi-directional MCP support and native deployment across a website widget, WhatsApp, Messenger, Slack, Microsoft Teams, Telegram and Twilio voice from one agent definition. The Autonomous Engine (LLMz) lets an agent plan and execute multi-step actions from a single instruction instead of a rigid flow, and a $25M Series B (June 2025, $120M valuation) on top of an earlier Series A signals real staying power for a company this size. The catch is the audience: Botpress is explicit that it isn't a zero-engineering tool — independent reviews describe needing at least one developer to wire up tool calls — and the LLM token cost ("AI Spend") bills separately from the plan fee, often exceeding it in production. Pick Botpress over Botsify if you want an open-source, code-level-customizable platform rather than a smaller no-code builder; pick Dify instead if the priority is a broader RAG/LLMOps workflow rather than conversational-agent channel deployment; pick a fully-managed option like Lindy if the goal is to avoid engineering involvement altogether.

## Strengths and weaknesses

- ✓ Genuinely open-source and self-hostable (MIT-licensed core, ~14.8k GitHub stars) rather than an open-core project that gates the useful parts behind a paid tier.
- ✓ Well-capitalized for a company this size — a $25M USD Series B in June 2025 at a $120M post-money valuation, on top of an earlier ~$15M Series A — with roughly 3,000 paying customers and revenue reported to be doubling every quarter.
- ✓ Bi-directional MCP support and native deployment across WhatsApp, Messenger, Slack, Teams, Telegram and voice channels from a single agent definition.
- ✗ Requires real engineering capacity — Botpress is positioned as a developer-first, code-level-customization platform rather than a fully no-code builder; wiring up tool calls and integrations generally needs at least one developer.
- ✗ The underlying LLM's token usage ("AI Spend") bills separately from the platform subscription and is often the larger cost in production — independent estimates put it at $200–$1,000+/month on top of the plan fee.
- ✗ Published pricing here is sourced from independent third-party reviews, not Botpress's own pricing page (which returned no accessible content at verification time); the two sources also disagree on the free tier's message cap (500/month vs. 5,000/month) — worth reconfirming directly with Botpress before budgeting.
- ⚠ Not a zero-engineering tool — independent reviews describe needing at least one developer to wire up tool calls and integrations.
- ⚠ LLM token cost ("AI Spend") is billed separately from the subscription and can exceed the plan fee in production use.

## Key features

- **Agent Studio** — A visual conversation-flow builder combined with code-level extensibility — custom JavaScript, API calls and webhooks — rather than a purely no-code canvas.
- **Autonomous Engine (LLMz)** — A custom inference engine that plans and executes multi-step, LLM-directed actions in place of a fixed decision tree.
- **Bi-directional MCP support** — The agent can both consume external Model Context Protocol servers and expose its own capabilities as one.
- **Knowledge Bases** — Indexes PDFs, websites, Notion, Google Drive and Confluence so answers are grounded in an organization's own content.
- **Multi-channel deployment** — Website widget, WhatsApp Business API, Messenger, Slack, Microsoft Teams, Telegram and Twilio voice from one agent definition.

## Use cases

- **Customer support deployed across channels** — A support team builds one agent in Studio, connects it to a knowledge base of help-center articles, and deploys it to a website widget, WhatsApp and Slack, with unresolved conversations routed to a human queue.
- **Self-hosted agent for data-residency requirements** — A regulated company self-hosts the open-source core on its own infrastructure to keep conversation data in-region, using the free MIT-licensed core rather than Botpress's managed cloud.
- **Internal workflow agent with tool access** — An engineering team wires the agent into internal databases and Salesforce via native API calls and MCP, letting it look up records and take action rather than only answering questions.

## Integrations

Salesforce, HubSpot (native CRM sync) · Notion, Jira, Calendly, Shopify, Stripe · 200+ apps via Zapier · WhatsApp Business API, Messenger, Slack, Microsoft Teams, Telegram, Twilio voice · Model Context Protocol (MCP) — bi-directional, as both client and server

## Sources

- https://github.com/botpress/botpress
- https://betakit.com/botpress-closes-34-million-cad-series-b-to-help-companies-build-and-deploy-ai-agents/
- https://www.lindy.ai/blog/botpress-pricing
- https://saleshive.com/vendors/botpress
- https://productgrowth.in/tools/customer-support/botpress/