Hexabot v3 - Open-Source AI Agent Framework

TL;DR

TL;DR: Hexabot v3 is an open-source Node.js automation platform that combines AI agents, YAML-based workflows, MCP tool support, and multi-channel deployment (chat, voice, API) in a single runtime.

Source and Accuracy Notes

⚠️ This section is MANDATORY. All links must be verified from actual source, not guessed.

• Project page: hexabot.ai ← MUST visit and verify
• Source repository: github.com/Hexastack/Hexabot ← MUST read README
• License: NOASSERTION (verified via GitHub API license.spdx_id returns null; LICENSE file not present in repo root)
• HN launch thread: news.ycombinator.com/item?id=41724497 — 19 points
• Source last checked: 2026-06-28 (commit main branch, pushed 2026-06-27)

What Is Hexabot?

Hexabot v3 is an open-source AI automation platform written in Node.js. It lets you build and deploy agentic workflows across multiple channels — chat, voice, and API — using YAML configuration files, a CLI tool, and first-class support for the Model Context Protocol (MCP).

From the README:

“Hexabot v3 is an automation platform with first-class AI capabilities, combining workflows, actions, and conversational channels in one runtime.”

The project targets Node.js ^24.17.0 and works with npm, pnpm, yarn, or bun as the package manager.

Setup Workflow

Step 1: Install the CLI

npm install -g @hexabot-ai/cli

Or run without a global install:

npx @hexabot-ai/cli --help

Step 2: Create a New Project

hexabot create my-project
cd my-project
hexabot dev

The hexabot create command auto-detects your package manager. Force one with --pm:

hexabot create my-project --pm npm

Step 3: Access the Local UI

After hexabot dev, the default local endpoints are:

• Admin UI: http://localhost:3000
• API: http://localhost:3000/api
• API docs: http://localhost:3000/docs

Step 4: Define a Workflow

Workflows are defined in YAML. A basic automation flow looks like:

name: welcome-flow
steps:
  - action: send-message
    message: "Hello! How can I help you today?"
  - action: wait-for-input
  - action: classify-intent
  - action: route-to-agent

Step 5: Connect MCP Tools

Hexabot v3 supports the Model Context Protocol, letting you wire in external tools and data sources:

hexabot tool add <tool-name>

Tools are then available to agents within workflow definitions.

Deeper Analysis

Architecture

Hexabot runs as a single monorepo with a CLI package, a core engine, and optional Docker services. The engine handles:

• Workflow engine — YAML-defined state machines with branching, loops, and error handling
• Channel adapters — Web, API, Discord, Slack, WhatsApp (extension-based)
• LLM integration — Claude, Gemini, Mistral, DeepSeek, Ollama (via standard API)
• Memory and RAG — Built-in vector storage for context retention

Core Capabilities

From the README’s capability list:

• YAML Workflows — Declarative automation definitions
• MCP Support — Model Context Protocol for external tool integration
• Multi-channel — Chat, voice, and API from one runtime
• RAG built-in — Retrieval-augmented generation with no extra setup
• Docker deployment — Full Docker Compose setup for production

Extension System

Hexabot has an extensions marketplace at hexabot.ai/extensions. Extensions cover additional channels, third-party integrations, and specialized AI providers.

Practical Evaluation Checklist

Prerequisites verified from source:

• Node.js ^24.17.0 (from package.json engines field)
• One of npm, pnpm, yarn, or bun
• Docker (optional, for containerized services)

Installation verified:

• npm install -g @hexabot-ai/cli (from npm registry listing)
• hexabot create auto-detects package manager (from README)
• Default ports: 3000 (Admin UI, API, docs) — verified from README

Features verified from README:

• YAML workflow definitions
• MCP tool support
• Multi-channel deployment (chat, voice, API)
• Built-in RAG
• Docker Compose production setup

Not verified (omitted from article):

• Specific API endpoints beyond default local ports
• Pricing for hosted/enterprise tiers (not mentioned in README)
• Exact list of supported LLM providers beyond named examples

Security Notes

• Hexabot v3 is self-hosted. You control your data and infrastructure.
• The platform handles conversation data and may integrate with external LLM APIs — review your LLM provider’s data handling policies.
• The GitHub repository is open-source. Audit the source before deploying in a security-sensitive environment.
• No authentication details are hardcoded in the default workflow templates.

FAQ

Q: Does Hexabot require an external LLM API key? A: Yes, Hexabot integrates with external LLM providers (Claude, Gemini, Mistral, etc.) via their APIs. You supply the API keys for the models you want to use. The platform also supports Ollama for fully local inference.

Q: Can I run Hexabot entirely offline? A: Yes — using Ollama as the LLM backend, Hexabot can run without any external API calls. Docker support is available for containerized self-hosted deployments.

Q: What channels does Hexabot support out of the box? A: The core platform supports Web chat, REST API, and has extension support for Discord, Slack, and WhatsApp. Additional channels are available via the extensions marketplace.

Q: How does Hexabot compare to similar tools like n8n or Langflow? A: Unlike n8n (general-purpose workflow automation) or Langflow (visual LangChain IDE), Hexabot is purpose-built for AI agentic workflows with first-class MCP support and conversational channels. It uses YAML rather than a visual editor.

Q: Is the code open source? A: Yes, the repository is public on GitHub. The license is listed as NOASSERTION in the GitHub API — verify the LICENSE file in the repo before legal/compliance use.

Conclusion

Hexabot v3 is a capable open-source option for teams that want to build multi-channel AI agent workflows without committing to a proprietary SaaS. The YAML-first approach makes automation readable and version-controllable, while MCP support future-proofs tool integrations against the evolving AI agent standard stack.

If you want to self-host AI automation with a low barrier to entry and a active open-source community (977 GitHub stars, Discord available), Hexabot v3 is worth evaluating. Start with npx @hexabot-ai/cli create my-project and have a running instance in under five minutes.