Intermediate

Botpress: building conversational agents for customer support

Botpress is a hosted platform for building conversational agents that answer on a website, WhatsApp, Slack or inside a helpdesk tool. You build them in a visual studio and can add JavaScript or TypeScript code when needed.

Data as of October 22, 2026

Key figures

DeveloperBotpress
TypeLow-code
LanguagesJavaScript, TypeScript
Open sourceNo
Licenseproprietary
First releaseNovember 2016
Suggested levelIntermediate
Use caseschatbot, RAG, automation
Self-hostableNo
Business modelFreemium
Cloud offeringYes
Multi-agentNot disclosed
MCP supportNot disclosed
Bring your own modelYes
Official siteOfficial site · Documentation · Repository

GitHub

Repositorybotpress/botpress
GitHub stars14,964
Forks2,296
Latest code activityOctober 22, 2026
Latest releasechat-2026-08-12_13-16-40 (August 12, 2026)

GitHub activity checked on October 22, 2026.

Built around conversation

Botpress focuses on one job: the chatbot and conversational agent, mostly for customer service. In Botpress Studio you connect nodes, including an “Autonomous Node” that lets a large language model decide where the conversation goes next. You can add knowledge bases so the agent answers from your own documents, a RAG-style approach, plus data tables. The “Execute Code” card runs JavaScript in the middle of a flow.

Listed integrations cover web chat, WhatsApp, Slack and tools such as Intercom, Zendesk, HubSpot and Salesforce. The company also offers Botpress Viber, which generates an agent from a plain-language description, and Botpress Desk, an AI-powered helpdesk. For developers, the ADK is a TypeScript framework for defining agents in code and deploying them to Botpress Cloud.

Open-source roots, cloud platform today

The Botpress GitHub repository dates back to November 2016 and is still MIT-licensed. Be aware, though, that the documentation states Botpress v12 and all self-hosted versions are officially sunset, and new users must go through Botpress Cloud. In practice, if you are looking for a chatbot tool to run on your own servers, Botpress no longer covers that need.

Pricing and lock-in

As of October 22, 2026, the pricing page lists, with annual billing, a Free plan at $0, Plus at $150 per month and Team at $750 per month, plus Enterprise on request (which adds the voice channel, among other things). The model is built on conversations: each plan includes a volume and an AI usage allowance, and extra conversations are billed in packs. For a high-traffic website, that volume, not the subscription, will drive the bill.

Lock-in works on two levels: your agents run on Botpress Cloud, and their logic lives in Botpress’s own format. Model choice, on the other hand, stays open.

Strengths, limits and alternatives

Botpress is strong for shipping a multichannel conversational agent quickly, wired into common support tools. Its limits: no self-hosting, a bill that tracks conversation volume, and a scope centered on dialogue rather than general automation.

For voice and chat agents, compare it with Voiceflow. If you want an open, self-hostable platform, Dify is worth a look. Official details: botpress.com/pricing.

Frameworks of the same type

Frequently asked questions

Is Botpress free?

Yes, there is a Free plan: as of October 22, 2026, it includes 100 conversations, 3 seats and 3 agents. The paid Plus and Team plans cost $150 and $750 per month respectively with annual billing. Enterprise is priced on request.

Is Botpress still open source?

The Botpress GitHub repository is MIT-licensed, but it mainly holds the SDK, the CLI and the integrations. The current platform, Botpress Cloud, is a proprietary service. Botpress v12 and all self-hosted versions have been officially sunset.

What counts as a conversation in Botpress billing?

Botpress counts a conversation once an end user sends at least two messages in a month. Each conversation includes a small AI usage allowance ($0.10), and extra conversations are sold in packs.

Can I choose the AI model in Botpress?

Yes. Studio lets you select a different large language model from the default one. Your AI usage cost then depends on the model you pick.

Data sources

Data as of October 22, 2026

Report an error

Does a figure, a date or a fact look wrong? Tell us: every report is checked.