Heym

Heym provides a visual, no‑code canvas for constructing AI‑driven automations that run on self‑hosted infrastructure. Users can describe a workflow in natural language and let the built‑in assistant generate the node graph, or manually connect a variety of pre‑defined nodes such as LLMs for text or vision, agent nodes with tool‑calling and Python execution, and Qdrant‑based retrieval‑augmented generation pipelines. The platform supports multi‑agent orchestration, persistent memory, automatic context compression, and human‑in‑the‑loop approval checkpoints, while offering guardrails and error‑handling features.

The system includes integrations with many services—Telegram, Slack, IMAP email, WebSocket streams, Redis, RabbitMQ, Grist, Google Sheets, BigQuery, and generic HTTP APIs—plus a Playwright browser automation node that can auto‑heal. It also connects to large‑language‑model providers such as OpenAI, Anthropic, Ollama, Cerebras, Google Gemini, and OpenRouter, and supports MCP client/server communication for external control.

Deployment is handled via Docker Compose or Kubernetes, giving operators full control over data and environment. The platform is source‑available, intended for experimental use, and targets developers who need a flexible, low‑code way to build and run complex AI workflows without writing extensive code.