Heym is an AI-native workflow automation platform designed for building, visualizing, and deploying intelligent AI workflows without requiring code. It targets engineers and power users seeking a robust, self-hostable solution where AI is the core execution model, offering benefits like streamlined development, enhanced observability, and flexible deployment.

How It Works

Heym employs an AI-first architecture, utilizing a visual drag-and-drop canvas powered by Vue.js and FastAPI. Its core approach centers on LLM and Agent nodes, enabling sophisticated multi-agent orchestration, RAG pipelines, and Human-in-the-Loop (HITL) capabilities. This design prioritizes AI as the fundamental execution paradigm, differentiating it from platforms that retrofitted AI capabilities onto existing automation frameworks.

Quick Start & Requirements

• Primary install/run: Docker is the recommended deployment method, with docker run commands provided for local setup and ./deploy.sh for production. Local development can be initiated with ./run.sh.
• Prerequisites: Bun ≥ 1.0, Python ≥ 3.11, UV, and Docker are required.
• Links: Product site: heym.run.

Highlighted Details

• AI Assistant: Generates workflows directly from natural language descriptions within the visual editor.
• AI Skill Builder: Enables creation and revision of agent skills using natural language, with live previews of SKILL.md and Python files.
• Multi-Agent Orchestration: Visually compose complex behaviors where one agent delegates tasks to named sub-agents or sub-workflows, supporting parallel execution.
• Built-in RAG: Integrates semantic search against managed Qdrant vector stores directly within workflows.
• Auto Heal: Playwright automation nodes automatically detect and fix broken selectors at runtime using AI.
• Observability: Comprehensive LLM Traces, Evals for AI test suites, and detailed Execution History provide deep insights.
• Portal: Transforms any workflow into a public or authenticated chat UI with streaming responses and file uploads.

Maintenance & Community

A Discord server is available for community interaction, questions, and updates. An enterprise offering provides commercial licensing, deployment assistance, and dedicated professional support.

Licensing & Compatibility

The project is licensed under the Commons Clause applied to the MIT License, making it source-available but not strictly OSI-approved open-source. Users are free to use, modify, and self-host, but commercial resale or offering as a paid service is prohibited without a separate commercial license.

Limitations & Caveats

The primary limitation stems from its licensing, which restricts commercial use without a paid license. While feature-rich, the "source-available" nature means it cannot be freely commercialized as a service, which may be a consideration for certain adoption scenarios.