Hexabot

Self-hostable visual workflow builder for shipping production conversational and agentic AI.

Freemium· Community free; Starter $19/mo; Pro $59/mo; Unlimited $149/mo; Enterprise customAgentsMulti-model

Best for

Pick Hexabot if you want a self-hostable visual agent builder with memory, RAG, and MCP that can graduate from prototype to governed production deployment.

Skip if

Skip it if you need a fully managed turnkey chatbot SaaS or an OSI-licensed open-source framework you can fork commercially without restrictions.

Hexabot is a source-available AI workflow automation platform for building conversational and operational agents you can actually run in production. Workflows are authored either visually on a canvas or as portable YAML with schema validation, then deployed across web, messaging, and custom channels with managed memory, RAG context, MCP tool bindings, and human-in-the-loop handoff baked in.

It sits between no-code chatbot builders and full code-only agent frameworks. The Fair Core License model means the codebase is inspectable and self-hostable for free, while paid tiers ($19/mo Starter through $149/mo Unlimited, plus Enterprise) unlock user seats, more workflows, and concurrent activations. That positioning makes it appealing to indie developers and agencies who want LangGraph-style flexibility without writing the orchestration scaffolding, and to regulated teams who need on-prem data residency.

The platform is LLM-agnostic, exposing action bindings to plug in different model providers, internal APIs, and MCP servers. The community tier (1 user, 3 workflows) is generous enough for serious evaluation, though the page itself is light on concrete model/integration documentation, pushing readers to the docs portal.

Editor's take

Hexabot lands in a useful gap: more opinionated than raw LangGraph, more inspectable than a closed SaaS like Voiceflow. The visual-plus-YAML duality is the right call for teams who need code review on their agents. We would like to see clearer docs on model providers and concrete integrations on the landing page itself.

— The AI Tool Bible editorial team