Hexabot

The editor is where you design conversation flows with a drag-and-drop interface. You can add text messages, quick replies, and carousels, then branch them based on what the user does. It keeps the logic readable for non-developers on the team, while the underlying runtime stays schema-first and code-driven.

Version 3 shifted Hexabot toward workflows defined in YAML with typed runtime contracts. That means the behavior of a bot lives in a version-controlled file you can review and diff, not in a vendor dashboard you can't export. For teams already treating automation as code, this is a meaningful change. No lock-in. Just files you own.

Actions define what a workflow actually does, with schema-validated inputs, outputs, and settings. A separate binding system lets you reuse capability and configuration across tasks, so you're not copy-pasting the same integration logic into every flow. It's the part that turns a chatbot into something that can get real work done. That's the jump from chat to action.

Hexabot treats channels as a core concept, so the same bot can run on web, mobile, and social messaging platforms. Conversation continuity carries across those surfaces. Does that matter? It does when customers start on one channel and finish on another. They don't have to repeat themselves.

The platform supports explicit memory definitions so the bot remembers context within a session, and a configurable RAG helper grounds AI answers in your CMS content. You can use the built-in database full-text helper, install a vector helper for SQLite or PostgreSQL, or write a custom one for another retrieval system.

Hexabot supports the Model Context Protocol for tool and context interoperability, and it connects to LLMs through Ollama, ChatGPT, Mistral, or Gemini. Swapping models doesn't mean rebuilding your workflows. Try one, then switch. That keeps you from getting locked into a single provider.

You extend Hexabot by developing and installing extensions from the Extension Library. That opens the door to text-to-action responses, third-party system integrations, and custom helpers. If your needs outgrow the built-in features, this is where you'd add them. Nothing's off the table.

Granular access control covers different parts of the system, and subscribers can be organized with labels and segments. The inbox provides a real-time window where conversations are monitored and handed over to a human agent when the bot can't finish the job.